close

Вход

Забыли?

вход по аккаунту

?

Microwave excited atomic spectral sources

код для вставкиСкачать
INFORMATION TO USERS
The most advanced technology has been used to photo­
graph and reproduce this manuscript from the microfilm
master. UMI films the original text directly from the copy
submitted. Thus, some dissertation copies are in typewriter
face, while others may be from a computer printer.
In the unlikely event that the author did not send UMI a
complete manuscript and there are missing pages, these will
be noted. Also, if unauthorized copyrighted material had to
be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are re­
produced by sectioning the original, beginning at the upper
left-hand corner and continuing from left to right in equal
sections with small overlaps. Each oversize page is available
as one exposure on a standard 35 mm slide or as a 17" x 23"
black and white photographic print for an additional charge.
Photographs included in the original manuscript have been
reproduced xerographically in this copy. 35 mm slides or
6" x 9" black and white photographic prints are available for
any photographs or illustrations appearing in this copy for
an additional charge. Contact UMI directly to order.
•UMI
Accessing the World's Information since 1938
300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA
Order Number 8821032
Microwave-induced thermoelastic tissue imaging
Chan, Karen Hing-Sheung, Ph.D.
University of Illinois at Chicago, Health Sciences Center, 1988
UMI
300 N. Zeeb Rd.
Ann Arbor, MI 48106
PLEASE NOTE:
In all cases this material has been filmed in the best possible way from the available copy.
Problems encountered with this document have been identified here with a check mark V .
1.
Glossy photographs or pages.
2.
Colored illustrations, paper or print
3.
Photographs with dark background
4.
Illustrations are poor copy
5.
Pages with black marks, not original copy
6.
Print shows through as there is text on both sides of page
7.
Indistinct, broken or small print on several pages
8.
Print exceeds margin requirements
9.
Tightly bound copy with print lost in spine
V
^
k**
10.
Computer printout pages with indistinct print,_JS1
11.
Page(s)
author.
lacking when material received, and not available from school or
12.
Page(s)
seem to be missing in numbering only as text follows.
13.
Two pages numbered
14.
Curling and wrinkled pages
15.
Dissertation contains pages with print at a slant, filmed as received
16.
Other
. Text follows.
UMI
MICROWAVE-INDUCED THERMOELASTIC TISSUE IMAGING
BY
KAREN HING-SHEUNG CHAN
.S., Virginia Polytechnic Institute And State University, 1979
M.S., The Ohio State University, 1981
THESIS
Submitted in partial fulfillment of the requirements
for the degree of Doctor of Philosophy in Bioengineering
in the Graduate College of the
University of Illinois at Chicago, 1988
Chicago, Illinois
THE UNIVERSITY OF ILLINOIS AT CHICAGO
The Graduate College
cc<*U^ J 7, /9fF
I hereby recommend that the thesis prepared under my supervision by
Karen Hing-Sheung Chan
Mlcrowave~Induced
entitled
Thermoelastic Tissue Imaging
be accepted in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
fa/
In Charge of Thesis
Recommendation concurred in
Committee
t
1W
(a—\.
*
THE
UNIVERSITY
Of
HUNOiS
PS
CHICAGO
D517 Rev. 1/87
Final Examination
to my parents
iii
ACKNOWLEDGEMENT
I would like to express my appreciation to Dr. James Lin for his
guidance, advice, and support for my study.
I am also deeply indebted to Dr. Jerry Sychra,
It has been a
great pleasure working with him and I am grateful for the many hours he
has devoted to me.
Thanks are also due to Dr. Earl Gose, Dr. Dan Pavel, Dr. Michael
Laszlo, and Dr. Thomas DeFanti for being in my committee, and Dr.
Thomas Foo for proof-reading my thesis.
This work was supported in part by the Office of Naval Research.
I would also like to acknowledge the Department of Radiology for their
financial support during my last year of my study.
I am also very grateful to Mr. George Ashman, Mr. Jenn-Lung Su,
Mr. Yu-Jin Wang and Mr. Man Keung Yuen for their technical support in
the laboratory.
Finally, I thank my parents, my family, Miss Tin Hung and my dear
friend Man Keung Yuen for their love, constant support, encouragement,
and patience throughout my studies.
KC
iv
TABLE OF CONTENTS
CHAPTER
1.
2.
PAGE
INTRODUCTION.
1
1.1 Background
1.1.1 , Ionizing Radiation
1.1.1.1
Film Radiography
1.1.1.2
Nuclear Medicine
1.1.1.3
X-ray Computerized Tomography
1.1.1.4
Positron Emission Tomography
1.1.2
Non-ionizing Radiation
1.1.2.1
Ultrasonography
1.1.2.2
Magnetic Resonance Tomography
1.1.2.3
Microwave-induced Thermoelastic Imaging. .
1.2 Objective?
1.3 Organization of This Thesis
2
3
3
4
5
6
7
7
10
11
16
17
THERMOELASTIC WAVES
18
2.1 Physical Properties of Biological Material
19
2.1.1
Microwave Properties
19
2.1.2
Thermal Properties
23
2.1.3
Acoustic Properties
27
2.2 Microwave Absorption Pattern And Induced Thermoelastic
Pressure Waves
31
3.
HARDWARE SYSTEM
37
3.1 Microwave Source And Its Control Circuit
3.2 Acoustic Wave Detection
3.3 Signal Processing And Data Conversion
3.3.1
Charge Amplifiers
3.3.2
Band-pass Filters
3.3.3
Amplifiers
3.3.4
Multiplexers
3.3.5
Data Conversion
3.4 Controlling Unit And its Interface
3.4.1
Microcomputer
3.4.2
Input/Output Expansion Board
3.5 Display
3.5.1
Color Monitor
3.5.2
Graphic Controller
37
43
45
45
46
50
54
54
59
59
62
66
66
66
v
J
TABLE OF CONTENTS (continued)
CHAPTER
4.
5.
6.
7.
PAGE
SOFTWARE SYSTEM
67
4.1 Data Acquisition
4.2 Subtraction And Normalization
4.3 Image Processing
4.3.1
Linear Image Manipulation
4.3.2
Non-linear Image Manipulation
4.3.3
Thresholding
4.4 Display
68
70
76
80
84
89
91
METHOD OF EXPERIMENT
93
5.1
5.2
93
94
Phantoms
Experimental Protocol
RESULTS AND DISCUSSION
100
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
6.10
100
103
107
120
123
123
127
127
129
130
The Microwave-induced Acoustic Source
A Model
Single Test Tubes
Biological Materials
Spatial Resolution
Diffraction Pattern
Time-Of-Flight Images
Attenuation Correction
False Triggering
Quality Control
CONCLUSION
131
BIBLIOGRAPHY
134
APPENDICES
140
Appendix A
Appendix B
Appendix C
140
149
153
VITA
170
vi
LIST OF TABLES
TABLE
PAGE
I
Microwave properties of biological materials
21
II
Thermal properties of biological materials
26
III
Parameters
28
IV
Acoustic properties of biological materials
30
V
Microwave generator product specifications
40
VI
I/O base address assignment
65
vii
LIST OF FIGURES
PAGE
1
2
3
4
5
A one-dimensional model of a plane wave impinging on
a tissue medium
22
Depth of electromagnetic wave's penetration in bio­
logical tissues as a function of frequency
24
Magnitude of the microwave power transmission at an
air-tissue interface
25
Frequency spectrum of acoustic wave induced by thermoelastic mechanism in water irradiated with 2450 MHz,
2 (j,s microwave pulse (theoretical calculation)....
33
Acoustic wave induced at the surface of the water tank
irradiated with 2450 MHz microwave pulses: tg = 2 ^s,
average power = 30 KW, and band-limited at 25 KHz and
250 KHz. (a) The received wave; (b) its frequency
spectrum
35
Acoustic frequency at maximum intensity as a function
of pulse width
36
7
A block diagram of the imaging system
38
8
A block diagram of a microwave generator
41
9
A block diagram of the external controlling circuit for
the pulse forming network
42
10
A thermoelastic wave received by the system
44
11
A charge amplifier
47
12
A cascaded low-pass, high-pass passive filter
49
13
A bode plot of the magnitude transfer gain of a filterand amplifier section
51
The phase characteristic of the filter, (a) input;
(b) output
52
15
An inverted amplifier
53
16
Functional diagram of a multiplexer
55
17
A schematic diagram of an analog-to-digital converter. . 56
6
14
viii
18
Timing diagram for the analog-to-digital converter. . .
58
19
A schematic diagram of a sample-and-hold
60
20
Functional diagram of I/O lines
64
21
An overall structure of ACQ4
69
22
A background image
71
23
An object image
73
24
A subtraction image
74
25
A plot of the first normalization technique as function
of the second normalization technique
77
A normalized image - using the first normalization
technique
78
A normalized image - using the second normalization
technique
79
26
27
28
Operators for smoothing filters
.82
29
A 3 x 3 low-pass operator and its processed image. ... 83
30
Operators for edge enhancement filters
85
31
A 3 x 3 high-pass operator and its processed image. . .
86
32
Grey-scale modification methods: (a) linear; (b) piecewise linear
88
33
Histogram equalization
90
34
Two dimensional bilinear interpolation
92
35
Simplified diagram of the microwave-induced acoustical
imaging apparatus
95
36
Experimental protocols (a) two test tubes; (b) three test
tubes
98
37
A typical amplitude background image
101
38
A time-of-flight background image
102
39
Non-uniform gain of processing channels
104
40
An "OBJECT" image: a 1.8 cm test tube of glass tube
ix
filled with muscle-equivalent material
A model
,
105
106
An image of a single test tube of 0.6 cm, filled with
muscle-equivalent material
108
An image of a single test tube of 0.9 cm, filled with
muscle-equivalent material
109
An image of a single test tube of 1.8 cm, filled with
muscle-equivalent material
110
45
The projection of a test tube
Ill
46
A thresholded image of a single test tube of 0.9 cm,
filled with muscle-equivalent material
112
A thresholded image of a single test tube of 1.8 cm,
filled with muscle-equivalent material
113
48
Histogram of muscle phantoms
115
49
Comparing the actual size of the test tubes to that
measured from the images
116
50
Errors from shadowing
117
51
(a) An ideal edge; (b) after moving average filtering.
118
52
An image of a 0.9 cm test tube placed (a) cm above the
transducer array; (b) on top of the transducer array. . 119
53
An image of two test tubes: (a) the top one filled with
muscle-equivalent material; (b) the bottom one filled with
glycol
121
54
An image of two test tubes: (a) the top one filled with
muscle-equivalent material; (b) the bottom one filled with
0.9 % saline
122
55
An image of three test tubes of 0.9 cm and 1.8 cm apart 124
56
An image of a single test tube filled with air
125
57
Histograms of images, (a) muscle phantom; (b) air. . .
126
58
A time-of-flight image of a 0.9 cm test tube filled with
muscle-equivalent material
128
59
Frequency spectrum of a microwave-induced acoustic wave
41
42
43
44
47
141
x
60
Frequency spectrum of a microwave-induced acoustic wave
142
61
Frequency spectrum of a microwave-induced acoustic wave
143
62
Frequency spectrum of a microwave-induced acoustic wave
144
63
Frequency spectrum of a microwave- induced acoustic wave
145
64
Frequency spectrum of a microwave-induced acoustic wave
146
65
Frequency spectrum of a microwave-induced acoustic wave
147
66
Frequency spectrum of a microwave-induced acoustic wave
148
67
A schematic diagram of the passive filter
150
68
A schematic diagram of the data acquisition system I .
151
69
A schematic diagram of the data acquisition system II.
152
xi
SUMMARY
A prototype microwave - induced thermoelastic tissue imaging system
was designed to acquire and process two-dimensional projections of
artificial objects.
A test object is immersed in a tank of water at whose surface
single microwave pulses of 2 /zs at a carrier frequency of 2450 MHz are
launched.
Thermoelastic
waves
induced
at
the
water
surface are
detected, on propagating through the object material, by a 20 x 20
piezoelectric transducer array at the bottom of the water tank.
received signals are then amplified and band-limited.
The
A computer-
controlled data acquisition system samples and converts the signals
into digital form.
The system produces images interactively through
the use of image processing techniques, which include subtraction,
normalization, convolution filtering, grey scale transformation, and
thresholding.
The processed images are displayed on a color monitor.
The circuit diagrams and the computer algorithms developed are shown.
Results are presented, encouraging the potential use of microwaveinduced thermoelastic waves as a possible medical imaging modality.
xii
CHAPTER 1
INTRODUCTION
In the last thirty years, particularly the last decade, advances
in the field of biomedical imaging has been rapid.
Imaging modalities
and protocols which only a short time ago were in the experimental
stage have become routine clinical procedures. In 1950s, for example,
nuclear medicine and ultrasonography were first introduced.
Today,
nuclear medicine is one of the major imaging modalities in diagnostic
radiology.
the U.S.
It performs over millions of functional studies yearly in
Real time ultrasonography is an important tool in cardiology
and obstetrics to image moving organs.
In 1970s, the development of X-
ray and magnetic resonance computed tomography is unquestionably the
most significant advancement in the diagnostic imaging field since the
discovery of X-ray (Kak, 1981; Brownell et al., 1982).
They have
become essential procedures of almost every diagnosis.
These
non-invasive
parameters
of
the
techniques
body
pathophysiological processes.
with
measure
different
differing
biophysical
sensitivity
for
Each technique has its own potential and
limitations, and no single method has prevailed to the exclusion of
others.
Nevertheless, the success of these techniques has prompted a
number of researchers to explore other physical and chemical properties
of
biological
materials
for
non-invasive
tissue
characterization.
Microwave-induced thermoelastic waves appear to possess some unique
features that may allow them to become as useful as these other methods
and to permit non-invasive imaging of tissue characteristics which are
1
2
not identifiable by other techniques (Olsen, 1982; Olsen and Lin,
1983).
Specifically, there is a direct relation between the pattern of
absorbed microwave pulses and the induced thermoelastic pressure waves
in biological tissues (Lin, 1978).
These different absorption patterns
are believed to be related to differing permittivities of tissues.
Thus, deep or surface acoustic illumination could be generated with a
selected microwave absorption.
The purpose of the thesis is to present
a system which uses pulsed microwave as an external source to obtain
two dimensional projections of some artificial tissue-simulated objects
and then processes and displays these images on a color monitor.
1.1 Background
In
all
medical
imaging
methods,
including
film
radiography,
nuclear medicine, ultrasonography, computerized tomography, positron
tomography, and magnetic resonance tomography, some form of energy is
transformed to a visual pattern.
The energy is either transmitted
through or reflected from structures deep inside the body.
Ideally, a
medical image is a physiological record which records either forms or
functions of certain organs.
Each of them possesses some properties
that make it uniquely useful and has its limitations in terms of
diagnostic capability and possible adverse effects.
information
on
the
above
mentioned
imaging
The background
modalities
and
the
literature review on the microwave-induced thermoelastic tissue imaging
will be presented in the following subsections.
They are grouped into
two categories depending on their capability of producing ionization,
that is, producing ions by ejecting orbital electrons from the atoms of
the material through which they travel.
The minimum photon energies
3
capable of causing ionization in biological material are between 10 eV
and 25 eV.
Therefore, 10 eV is generally considered as the lower limit
for ionization in biological systems (Lin, 1978).
1.1.1 Ionizing Radiation
1.1.1.1 Film Radiography
Film
radiography
diagnostic radiology.
is
the
simplest
and
most common
method
in
It consists of irradiating the relevant part of
the body with an X-ray source and allowing the unabsorbed X-rays to
fall on an X-ray sensitive fluorescent screen.
off by the screen
The visible light given
blackens a sheet of light-sensitive film placed in
close contact with the screen, thereby producing a two-dimensional
attenuation coefficient distribution function of a three-dimensional
structure.
This attenuation coefficient has a linear relationship with
the tissue density.
The frequency of most diagnostic x-ray is between
3 x 10" Hz and 3 x 10^ Hz. Therefore, film radiography is capable of
very fine spatial resolution in the range of 0.1-0.5 mm and is very
effective for imaging certain parts of the body with high contrast,
such as lungs and bones.
However, it is limited in its capability to
discriminate soft tissue organs.
To overcome this difficulty, high
contrast agents are introduced intravenously to enhance the visibility
of certain organs.
Side effects, such as nausea, are reported.
simple,
and
efficient,
economic
numerous books in medical physics.
procedure
is
well
This
documented
by
Both technical and clinical aspects
of X-ray film radiography can be found in Christensen et al. (1972),
Hendee (1979) and Macovski (1987)
4
1.1.1.2 Nuclear Medicine
Nuclear medicine has diagnostic values in assessing functions of
organs , coronary artery diseases, and in detecting cancers that have
spread to bone.
A small amount of radioactive tracer is injected
intravenously to the patient.
Depending on the radiopharmaceutical
used, the isotope will distribute selectively into a particular organ
system.
The most common isotope used is metastable technetium-99,
which emits a single photon energy of 140 KeV; other isotopes emit
photons with energy ranging from a few tens to a few hundreds of KeV.
The steady state or dynamic distribution of the injected isotope is
determined by a gamma camera, an area imager that produces a projection
image.
which
The gamma camera uses a collimator to determine the line along
the
multiplier
radioactive
array
to
decay
occurred,
determine
the
emitted
computer to display the resulting images.
mm can be achieved.
a
scintillator
photon's
and
energy
photo
and a
An overall resolution of 3-6
Basic information on routine diagnosis in nuclear
medicine can be found in Rassow (1980).
As in film radiography,
nuclear medicine suffers the shortcoming in which a three-dimensional
object
is
represented
as
a
two-dimensional
display.
Acquiring
complementary views at different angles is used clinically to give the
third dimensional information.
In addition,
image
processing
methods have
been proposed
to
generate functional images for data reduction, which compresses all the
information found in the sequence of images into a smaller number of
images, and for separation of " physiological factors" from overlapping
anatomical
structures.
For
examples, conventional functional (or
parametric) images, such as stroke volume in cardiac studies, have been
introduced into the United States since 1976; phase and amplitude
images, which are derived from the first harmonic of time activity
curves of a series of dynamic images (Pavel et al., 1983, Byrom et al.,
1981) come after; factor analysis images, which are estimates of
different physiological components from time activities curves, are
under investigation (Cavailloles et al., 1984).
1.1.1.3 X-rav Computerized Tomography
Computerized tomography (CT) is a radiological method of producing
thin cross-sectional images of the human body.
the
collection
of
a
large
number
transmission measurements, from which
mathematically reconstructed.
of
a
The technique involves
finely
collimated
tomographic
x-ray
image can be
In order to reduce data acquisition
time, a number of CT scanning mechanisms have been devised to make
multiple transmission measurements from a large number of angles.
See
Newton (1981) and Bates et al. (1983) for description of scanning
mechanisms for different generation scanners.
In current clinical
setting, a CT scanner usually involves a thin collimated fan beam of xray and an array of detector elements.
Each detector element produces
a charge signal proportional
absorbed x-ray energy, which
to
the
reflects the attenuation along the ray path between the element and the
x-ray source.
The depth
information needed
to
produce a cross-
sectional image is achieved by either rotating both the x-ray source
and the detector (3rd generation), or only the x-ray source with a
fixed detector (4th generation) about
the body axis while taking
projections of the same section from a large number of different
6
angles.
Data accumulation typically takes 1-2 seconds.
The data for
each projection are digitized and can be corrected for polychromaticity
of
the
x-ray
scattering.
source,
cross-talk
between
detector
elements
and
The digital information is then used to reconstruct a two-
dimensional map of the local attenuation coefficient.
A number of
approximation methods have been devised to provide a solution to the
problem
of
image
measurements.
reconstruction
from
these
half-million
or
more
Most of the commercial scanners use a variation of a
technique referred to as the "filtered back projection".
For detail
mathematical derivation of this reconstruction process, see Herman
(1980).
Finally, a true cross sectional image resulted from this
reconstruction process is displayed with an appropriate grey scale and
window.
Computed tomography is capable of providing spatial resolution
in the range of 1 to 1.5 mm and slice thickness in the range of 1-15
mm.
CT has been widely used
invention in the early 70's.
in diagnostic radiology since
its
It is proven to be useful in determining
the site, type, and extent of disease, guiding biopsies, planning and
monitoring therapy, and following up treatment.
1.1.1.4 Positron Emission Tomography
Positron emission tomography or PET technique has the potential
for measurement of regional functions such as glucose, fatty acid,
amino acid, and other substrate metabolism, receptor concentration in
the body, and blood flow (Brownell et al. , 1982).
nuclear
medicine
intravenously
to
procedures,
the
patient.
radiopharmaceutical
These
Similar to the
is
injected
radiopharmaceutical
emits
positrons (positive electrons), each of which, after travelling a
7
distance from the nucleus, interacts with a negative electron and
results
in
producing
two
photons
emitted
in opposite directions.
Detectors made of scintillation crystals are mounted in one or more
rings surrounding the patient to determine the line along which a
disintegration occurs.
Most PET imaging devices use time coincidence
to determine the line of position along which the positron emitter
decays.
In addition, it also uses the small differential in time of
arrival of the annihilation photons to estimate the position of the
source along this line of flight.
Most current instruments are capable
of producing images with a resolution of 10 to 20 mm FWHM.
Full width
at half maximum (FWHM) is the smallest separation at which two signal
sources can be placed and still be detected.
The most important
limiting factors to the resolution of positron imaging system are the
distance the positron travels through the tissue before annihilation,
o
the
angular
deviation from
180
for
the
two
photons
emitted
on
annihilation, finite detector dimensions, and statistical aspects of
reconstruction.
very
useful
In spite of its low resolution, PET has found to be
in
measuring
blood
flow
in
tissue,
consumption and glucose metabolism in the brain.
research
tool
for
the
investigation
on
diseases
determining
O2
It is a valuable
such
as
aging,
schizophrenia, atherosclerosis, and cancer (Brownell et al., 1982).
1.1.2 Non-ionizing Radiation
1.1.2.1 Ultrasonography
Ultrasonography has a distinct advantage over other techniques, in
that it involves acoustic waves, a form of radiation that appears to
have no significant side effects.
It is widely used in cardiology and
8
obstetrics
to
image moving organs or fetuses.
Short bursts of
ultrasonic energy are transmitted into the body at a resonant frequency
between 1 and 10 MHz, depending on application.
The pulse echo or
reflective approach has been most widely used.
In a reflectance
ultrasonogram, the system measures the returning echo time delay and
intensity, which give information about the acoustic impedance and the
depth of tissue interfaces.
The sound pulse continues to lose energy
as it transverses tissues.
Moreover, strong boundaries, like those
between tissue and bone, or tissue and air, reflect most of the
incident acoustic energy and prevent acquisition of data from more
distant structures.
Nonetheless, resolution along the axis of the
acoustic beam is about 0.5 mm and less than about 1 mm perpendicular to
that axis.
It is also used to measure blood flow, which in turn
evaluates the extent of narrowing of blood vessels by atherosclerosis.
It is invaluable for two main reasons.
biological side effects.
Second, it contributes useful diagnostic
information on soft tissue.
acoustic waves penetrate
rapidly in air.
First, it has no significant
However, it suffers the drawback that
poorly
at bony
structures and attenuate
It is not usually of much value for conditions
affecting the brain or lung.
Since the late 1960s, the possibility of using an alternative
approach, that of utilizing transmission ultrasonic imaging, has long
been discussed, but it has never become a routine clinical procedure
(Schuy, 1982).
Several researchers are attempting to develop practical
device able to display in real time a dynamic ultrasonic image.
In the
early 1970s, Holbrooke et al. (1974) was the first group gone through
9
preliminary clinical evaluation in ultrasound imaging.
They used a
focused liquid surface acoustical holography system to investigate
imaging pregnant uterus and developing fetus.
They have found that a
large amount of detail can be seen in fetal imaging with an excellent
correspondence with the actual morphology.
At about the same time,
Green and his group developed a new transmission ultrasound camera
system using a piezoelectric array for detection
Marich et al., 1975; and Zatz, 1975).
(Zatz et al., 1975;
They reported their in-vitro
studies on biological structures and preliminary studies on normal
adults
Since then, reports on using transmission ultrasonic imaging
for the lungs (Meltzer et al., 1981), the hip joint in the immature dog
(Weigel et al., 1983) and the upper limb (Hentz et al., 1984) have
promising
results.
developing
and
In
addition,
experimenting
investigators
computed
have
transmission
also
been
ultrasound
tomography (Carson et al., 1977; Greenleaf and Bahn, 1981; Greenleaf,
1982a, Greenleaf, 1982b, Greenleaf et al., 1982a, and Greenleaf et al.,
1982b) in which the received signals were processed for arrival time
and for changes in amplitude.
The measured values for arrival time and
attenuation were used in a convolution-back projection reconstruction
algorithm to obtain estimates of the two-dimensional distribution of
acoustic speed, also known as time-of-flight images and attenuation
within
the
scanned
planes.
The
effects
of
refraction
and
of
diffraction cause aberrations in the images, resulting in errors both
in geometry and in magnitude of the reconstructed values.
of
these
tomography'.
effects
lead
to
the
investigation
on
Correction
'diffraction
10
1.1.2.2 Magnetic Resonance Tomography
Magnetic resonance (MR) imaging is a new and very promising non­
ionizing imaging modality.
normal
and
cancer
It has potential for discriminating between
tissues, for
measuring
blood flow without any
contrast agent, and for generating images in response to intrinsic
tissue parameters, which
represent gross chemical characteristics of
the body (Brownell et al. , 1982).
Magnetic resonance projections can
be obtained in many different ways.
In general, a static field is
first applied to setup magnetic alignment of nuclei with an odd number
of protons.
An alternating magnetic field then applied by a nearby
coil causes nuclei to precess about the axis of the static field.
This
results in absorption of energy, which is then emitted when the nuclei
return to the equilibrium state.
The absorption and emission of energy
take place at the resonance frequency governed by the famous Larmor
theorem. Spatial information is obtained either by varying the field in
a known manner or by oscillating the magnetic field gradient
different directions.
in
See James (1983) and Shaw (1986) for detail
discussion of these methods.
Reconstruction algorithms similar to
those produce x-ray CT images may be used to deduce two-dimensional
cross-sectional images.
The MR images are multiparametric images
reflecting
of
distribution
proton
density
and
of
relaxation time, T1 and spin-spin relaxation time, T2.
spin-lattice
They are
believed to carry information about disease states of tissues.
Since
MR was used to form images in 1972, many efforts have been directed to
improving
data
acquisition
time
algorithms to enhance these images.
and
applying
image
processing
11
1.1.2.3 Microwave-induced Thermoelastic Imaging
Microwave-induced thermoelastic (or thermoacoustic) imaging with
application in biomedical area is still in its early experimental
stage.
Many
investigators have studied the feasibility of using
thermoelastic waves as a potential medical imaging modality.
Each of
them used a slightly different approach in terms of their experimental
set-ups.
below.
Their approaches and preliminary results are summarized
The
thermoelastic
literature
waves
in
review
biological
only
includes
tissue
applications
imaging.
For
of
other
applications such as nondestructive evaluation of inanimate material,
see Hutchins and Tam (1986a) and Hutchins and Tam (1986b).
Olsen and Hammer (1980) used a Navy type E-27 standard hydrophone
to measure the acoustic waves generated in a mass of tissue equivalent
material by the absorption of 200 KW, 0.5 fis pulsed microwave energy.
In that.study, a radar transmitter of 5.655 GHz was used to irradiate
the mass at close range, and the hydrophone response showed a well
defined acoustical pulse generated from the region where the microwave
energy
was
absorbed.
Su (1985) and
Su
and
Lin (1987) further
investigated thermoelastic signatures of tissue phantoms.
A 20 KW,
2450 MHz, 2 fj,s pulsed microwave source and a hydrophone transducer (0.7
cm in diameter) were used for their experiments.
They developed a
pattern extraction algorithm to analyze the acoustic waves pattern
generated from biological simulated phantoms contained in test tubes of
various sizes.
They suggested that thermoelastic waveform would be
proportional to the size of the test tube and dependent on the type of
solution within the test tube.
12
Olsen and Lin (1981) reported measurement of microwave-induced
acoustic pressure in spherical model of human and animal brains.
A 4-
kilowatt peak power, 10 /is microwave pulse at 1.1 GHz was applied to
the spherical models using an open-ended WR-650 waveguide of 46 cm
long.
The acoustic waves were received by a 1-cm-in-diameter spherical
hydrophone made of barium titanate piezoelectric element.
Similar
experiments have been done on animal brains, such as rats, guinea pigs
and cats (Lin et al. , 1982; Olsen and Lin, 1983).
results obtained agreed with
Their experimental
those, predicted by the thermoelastic
theory of interaction in regard to pulse width and the frequency of
impinging microwaves, pattern of absorbed microwaves energy, frequency
vibration, and threshold of sensation presented in Lin (1977a), Lin
(1977b) and Lin (1978).
In 1982, Caspers and Conway also reported their measurement of the
spatial power density distribution in lossy inhomogeneous material
during RF- and microwave hyperthermia (Caspers and Conway, 1982).
They
used four open-ended (semi-rigid) coaxial cables, immersed in a water
bath and operated at 9 GHz, 1 (is pulse width and 25 W peak power to
realize thermoacoustic point source excitation.
Response of a 50 KHz
and
experimental
500
KHz
microphone
was
recorded.
The
indicated that power density reconstruction of
results
a 50 KHz response would
be very difficult due to reflections from the boundaries and the poor
time resolution.
discernible.
Moreover, the four "point sources" were clearly
The results also confirmed the optimum microwave pulse
width for maximum acoustic signal predicted by Borth and Cain (1977).
13
Olsen
(1982)
dimensional form.
made
of
first
thermoelastic
images
in
two-
They used an 8x8 piezoelectric transducer array
lead zirconate
detector.
reported
titanate
material
as
their
acoustic wave
Each transducer element measured 1.9 cm in diameter and had
free field voltage sensitivity of -201 ± 0.1 dB at 100 KHz.
pulses of 14.5 and 20 ps at a carrier frequency of 1.1 GHz
Single
with peak
power of 4 KW were delivered to a tank of water through a standard Lband waveguide.
submerged
at
Output waveforms from the transducer array, which was
the bottom of the water
tank, were observed on an
oscilloscope serially and the peak-to-peak amplitude of the first wave
to
reach each of the
transducer element was recorded. Data were
collected first in the absence of a phantom as reference and then with
the phantom placed atop the array.
A mean amplitude of response for
the array was calculated from the reference data as an averaging
factor.
Amplitudes from the second set of data were subtracted from
those originally obtained pixel by pixel and
multiplied
by
the
averaging
factor.
The
each difference was
resulting
values
were
represented in a two-dimensional matrix of darkened circles, whose
radii were proportional to the net value difference at that point.
Crude images showing the overall features of the respective phantoms
were clearly seen.
A year after, Olsen and Lin (1983) attempted to
improve the image resolution of their system by using a higher power
microwave source operating at a higher frequency and a shorter pulse
width, and a higher
resolution transducer array.
In this study, a
military radar transmitter at 5.66 GHz was used to generate 200 KW, 2
/is microwave pulses.
These pulses were directed by a standard horn
14
located just above a tank of water.
A 20 x 20 piezoelectric transducer
array made of the same material and with the similar free field voltage
sensitivity as that reported in Olsen (1982) was used.
transducer elements measures 0.9 cm in diameter.
Each of these
A model of a human
hand filled with tissue-equivalent material was used as their phantom.
The experimental procedure was repeated as in Olsen (1982).
Data were
then keyed into a PDP11/23 minicomputer for further processing and
resulted images were displayed on a graphic terminal.
Then, Lin and
Chan (1984) presented a system design automating the data acquisition
in the previously reported experimental procedures (Olsen, 1982; Olsen
and Lin, 1983).
A hybrid parallel/serial design for dividing the 20 x
20 piezoelectric transducer array into segments and collecting data
from each segment sequentially was suggested.
The signal from each
transducer was amplified and band-limited before digitization.
bit analog-to-digital converter was used to obtain 256 levels.
An 8The
above reports all measured the attenuation of thermoelastic waves
generated by a single microwave pulse at and near the surface of a
phantom as they propagated through the tissue.
Their preliminary
results led to the present research.
At about the same time Olsen and Lin's group started investigating
microwave-induced thermoelastic tissue imaging, Bowen et al. (1981)
suggested the use of radiation-induced thermoacoustic waves for soft
tissue imaging.
In that study, 75 V to 1.2 KV electrical current
pulses of 0.4 ns duration was applied across a pair of copper plates to
a layer of soft tissue phantom and the generated acoustic signals were
detected by a 0.5 MHz 19 mm diameter unfocused transducer located
15
perpendicular
to
the direction of the applied electrical Tcurrent.
Typically, 10 to 10^ waveforms were digitized at 0.1
interval and
averaged by a PDPll/34 minicomputer before displaying them on a graphic
terminal.
The properties of the induced acoustic signals agreed with
the predictions of the theory of thermoacoustic emission reported in
Bowen (1981). Bowen et al. (1983a) and Bowen et al. (1983b) reported an
observation of acoustic signals from a phantom in an 18 MeV electron
beam for cancer therapy.
In this study, a 18 MeV therapeutic electron
beam was collimated to a 2.5 cm x 2.5 cm square cross section at the
entrance to the phantom.
The machine was operated at 180 rad/min
measured at 100 cm distance, 180 pulse/s repetition rate, and 1.5 (j,s
beam pulse width.
A 19 mm diameter, 0.5 MHz transducer and a signal-
averaging system, similar to the arrangement reported in Bowen et al.
(1981), were used.
Again, the experimental results agreed with the
theoretical prediction in Bowen (1981).
Nasoni et al. (1984) reported
the measurement of thermoacoustic waves induced at the interfaces of
different materials placed in a small water tank, itself inside a
microwave cavity.
A pulsed 2450 MHz magnetron was used to deliver
microwave energy to the cavity.
It was operated at 100 ms intervals
and these pulses were 0.5 /zs at half amplitude. The acoustic waves were
received by a 1 MHz resonant PZT5 transducer, sampled at 20 MHz, 100
MHz and 5 MHz sampling frequencies and then averaged over 1 to 64000
samples.
The processing was done in an IBM/XT computer.
The results
also agreed with the theoretical calculation.
The
above
theoretical calculation and experimental results of the
reports
suggested
that
thermoacoustic
imaging
may
permit
16
identification of tissue characteristics which are not sensed by other
means,
and
verification
positioning
of
radiation
of radiation dosage.
therapy
treatment
plans
in
The main difference between this
approach and the previous one is that this applied repeated pulses with
lower energy per pulse to the phantom and employed signal-averaging
method to the detected acoustic waves to improve the signal-to-noise
ratio.
1.2 Objective
The objective of this work is to design, build, and test a
prototype
system
that can obtain two-dimensional microwave-induced
thermoelastic wave projections of some artificial tissue simulated
objects.
For this study, only surface microwave absorption pattern is
considered, that is, thermoelastic waves are generated at the surface
of the tissue and then received by a 20 x 20 piezoelectric transducer
array underneath the tissue.
obtained sequentially.
Sixteen segments of one whole image are
It is assumed that the objects being scanned
are motionless throughout the data acquisition period.
The hardware
design is limited to obtain two dimensional images from a 20 x 20
piezoelectric transducer array with frequency response from 25 KHz to
250 KHz and with the above mentioned scanning method.
For the software
design, effort has been devoted to accommodate future expansion of the
system, or to apply the system to images from other imaging modalities.
The existing protocol is the first attempt to automate this imaging
system.
Many
implementation.
limitations
are
imposed
in order
to simplify
the
However, it is hoped that this prototype will be able
17
to allow researchers to further investigate the feasibility of using
microwave-induced thermoelastic waves as an imaging modality.
1.3 Organization Of This Thesis
The background information on six major medical imaging modalities
and literature review on microwave-induced thermoelastic imaging are
presented in chapter one.
In chapter two, the basic microwave-induced
thermoelastic wave characteristics and their frequency spectrums are
described.
Theoretical calculation and experimental data are included
for comparison.
Relationship between microwave source parameters and
the induced thermoelastic pressure waves in biological tissue can be
inferred.
Introduced in chapter three and four are the hardware design, the
implementation of data acquisition and image processing routines for
the thermoelastic tissue imaging system.
Several examples are given to
demonstrate the performance of the system.
phantoms and the testing protocols.
Chapter five describes the
They are specially designed to
test the spatial resolution and the tissue identification capability of
the system.
Finally,
in
chapter
six,
data
collected
by
processed with differing image processing routines.
limitations
of
the
system
are
suggestions conclude this thesis.
then
discussed.
the
system
are
Advantages and
Further research
CHAPTER 2
THERMOELASTIC WAVES
As early as the beginning of 1960, it was reported that human
beings could hear pulse-modulated low average power density microwave
transmitted through air.
effect
was
not
The mechanism responsible for this auditory
clearly
understood.
Throughout
these
years,
investigators have attempted to account for the effect from physical
and physiological consideration.
Lin (1978) described some of the
suggestions on the possible mechanisms involved.
The thermoelastic
converting mechanism first suggested by Foster and Finch (1974), after
examining
the
information on the conversion of electromagnetic, to
acoustic energy by surface heating (Gournay, 1966; White, 1963) has
been viewed as the most probable cause of microwave-induced auditory
sensation in mammals.
The energy conversion aspect was confirmed by
experimental reports from Sharp et al. (1974), Guy et al. (1975), Lin
(1976) and Lin (1978).
According to the thermoelastic theory, a
portion of the incident radiation is absorbed by
the
tissue and
converted into heat which generates a temperature gradient normal to
the surface.
As a result of thermal expansion occurring within a few
microseconds, this temperature produces strains in the tissue material
and leads to
generation of stress waves which propagate away from the
surface (Lin, 1978).
In this chapter, the parameters of the microwave source and the
properties of biological tissues directly related to the generation of
acoustic
waves
by
pulsed
microwave
18
energy
will
be
described.
19
Theoretical and experimental data on microwave-induced thermoelastic
wave frequency spectrum will be included for comparison.
between microwave
source
parameters
and the
Relationship
induced thermoelastic
pressure waves in biological tissue and the feasibility of using this
imaging modality will be discussed.
2.1 Physical Properties of Biological Material
The
microwave - induced
propagation
in
biological
thermoelastic
tissue
medium
waves
involve
generation
three
and
physical
processes: (1) microwave absorption; (2) conversion from microwave
energy to heat energy; and (3) stress waves generation.
Obviously, the
physical properties involve in these energy conversion processes are
the microwave, thermal and acoustic properties of biological tissues.
In this section, these three physical properties will be reviewed for
better understanding of this acoustic wave generation mechanism.
2.1.1 Microwave Properties
The
microwave
properties
of
biological
characterized by the dielectric constant,
conductivity, a.
the
amount
of
er
and
structures
the
are
effective
These two parameters together are known to determine
energy
transmitted
into
and
absorbed
by
tissues.
Measurements of them for - various tissues at different temperature in
the radio-frequency and microwave frequency range have been extensively
studied by Schwan and others.
They are found to be temperature
sensitive, highly frequency dependent and are largely determined by the
relaxation properties of biological membranes and tissue water content.
Michaleson and Lin (Chapter 4, 1987) described the relaxation process,
and methods of permittivity measurement, and summarized the measured
20
tissue
dielectric
constants
and
conductivities
as
a
function
of
frequency and temperature.
In order
to
simplify
the
discussion, biological material is
classified into three major groups according to their water content.
The first group is of high water content (90% or more) such as blood
and cerebrospinal fluids.
The second is made up of moderate water
content (less than 80%) such as skin, muscle, brain and internal
organs.
The third is of tissue with low water content (about 40%) and
consists of bone and fat (Lin, 1978).
data for each group.
TABLE I shows the representative
It is seen that the dielectric constant decreases
with increasing frequency while
the conductivity increases in the
frequency range from 100 to 10000 MHz.
Michaleson and Lin (1987) also
pointed out that dielectric constants of the in-vitro data of high
water content biological materials are in close agreement with in-vivo
measurements; and those of moderate and low water content have lower
values than the in vivo measurements.
These differences in dielectric
constants can be accounted for by the difference in water content due
to in-vitro and in-vivo
measurement conditions.
For a simple one-dimensional model in which a plane wave impinges
normally on a semi-infinite homogeneous tissue material as in Fig. 1,
the power density at a distance x from the surface is given by
I = I()e~^amX
= 0
> f°r
0 < t < t0
; elsewhere,
(2.1)
where Iq is the power density at the surface; tg is the pulse width;
and am is the attenuation coefficient.
The attenuation coefficient,
21
TABLE I
MICROWAVE PROPERTIES OF BIOLOGICAL MATERIALS
(borrowed from Lin, 1978 and Michaelson and Lin, 1987)
Frequency
MHz
0.9 % Saline
er
a
Muscle
er
a
Fatty Tissue
er
a
100
78
1.,67
72
0,,889
7.,45
19.1-75.9
200
78
1..68
56
1.,280
5.,95
25.8-94.2
400
74
1.,72
53
.430
1,
5..60
37.9- 118
700
77
1.,85
52
1,,540
5.,60
49.8- 138
1000
77
1.,88
5.,60
55.6- 147
-
47
2,,210
5.,50
96.4- 213
...
46
2,.260
5,,50
110 - 234
...
44
3.920
5,.50
162 - 309
...
40
7.650
4,
.70
255 - 431
7.65
40
10.300
.50
4,
324 - 549
2500
-
-
-
3000
-
-
- •
5000
8000
10000
-
-
66
- •
(Dielectric constant = er)
(Conductivity = a, mho/m)
22
Thermoelastic Tissue Imaging
TISSUE
AIR
U o / £ o /Cr o
::d£:0'ere0'a0
MBHI
DIRECTION OF PROPAGATION
A Plane Wave Impinging Normally
on a Semi-infinite Tissue Medium
Fig. 1 A one-dimensional model of a plane wave impinging on a
tissue medium (borrowed from Lin, 1978)
23
which is a measure of the rate of energy loss in the medium, is related
to the er and a by the following equation (Michaleson and Lin, 1987):
,."2,
+
c
2
)1/2 - 1U1'2
o/e^eo
(2.2)
Figure 2 graphically depicts some of the data reported in Michaleson
and Lin (Table 5.2, 1987).
The amount of incident microwave energy transmitted across the
interface of two biological media is a function of the dielectric
constants of both media, typically air and tissue.
Figure 3 is a plot
of magnitude of the microwave power transmission at an air-tissue
interface reported in Lin (Table XI, 1978).
From the above discussion, it is evident that transmitted energy
is strongly frequency dependent.
For higher frequencies, most energy
is absorbed at the surface and the wave can generate a surface acoustic
illumination.
penetrate
deep
For
lower
inside
frequencies,
the
tissue
and
the
microwave
allow
a
energy
deep
can
acoustic
illumination.
2.1.2 Thermal Properties
The conversion of microwave energy
transient
and
steady
state
temperature
into heat energy and the
distribution
rely
on
the
specific heat and thermal conductivity of the material exposed to
microwave irradiation.
The specific heat of a material is a measure of
the amount of energy required to raise, for one gram of the material,
one degree in temperature, while the thermal conductivity measures how
fast the heat energy is distributed.
TABLE II represents some of the
24
Depth Of Penetration
Saline
• Blood
Depth, cm
• muscle
O lun
A fat
\
2000
4000
6000
8000
10000
Frequency, MHz
Fig. 2 Depth of electromagnetic wave's penetration in biological
tissues as a function of frequency (borrowed from
Michaelson and Lin, 1987).
'
25
Power Transmission Coefficient
Coefficient
0.45
0.40 -
/
•
0.35-
I
0.30 - l
m
0.250.20
2000
4000
6000
8000
10000
Frequency, MHz
Fig. 3 Magnitude of the microwave power transmission at an airtissue interface
26
TABLE II
THERMAL PROPERTIES OF BIOLOGICAL MATERIALS
(borrowed from Lin, p. Ill, 1978)
Material
Thermal
Conductivity
cal/m sec°C
Specific
Heat
cal/g°C
Coefficient
of Thermal
Expansion
10"5(°C)-1
Distilled Water
0.15
0.998
6.9
Brain
0.126
0.88
4.14
Muscle
0.122
0.75
4.14
Fat
0.0525
0.62
2.76
Bone
0.35
0.49
2.76
27
data Lin (1978) summarized after examining the available information
from Chato and others.
The approximate temperature distribution, V(x,t) inside the medium
is given by Lin (1978) as
V(x,t)= 2amI0te"2Q!mx/^s
(2.3)
Parameters are explained in TABLE III.
In biological material, the
cooling curve is a slowly varying time function in the order of
milliseconds.
Therefore, when the microwave energy is deposited into
the tissue in a very short period of time (microseconds) , an extremely
rapid temperature rise and a large temperature gradient occur in the
tissue.
This temperature rise in turn produces a stress wave inside
the tissue.
This stress waves production is determined by the Young's
modules and coefficient of thermal expansion of the material.
TABLE II
also lists the coefficient of thermal expansion of various biological
tissues.
2.1.3 Acoustic Properties
The propagation of sound energy in a medium is governed by the
sound velocity and
experimental
the attenuation of the
techniques
for
the
medium.
determination
of
Many diverse
the
ultrasound
propagation parameters of biological tissues have been published and
reviewed by Wells (1977).
Ultrasound velocities in different tissues, with the exception of
those for bone and lung, are close to that in water.
In general,
1540 m/s is taken as the average velocity for body tissue (Christensen
et al., 1972).
In the medical ultrasound range, the velocity of
28
TABLE III
PARAMETERS
Parameters
am or a
aa
Iq
C
p
S
T
x
t
o>
f
a
sq
er
7
p(x,t)
u(x,t)
Explanation
microwave attenuation coefficient
acoustic attenuation coefficient
microwave intensity
velocity of elastic wave propagation for the
material
density of the material
specific heat of material
pulse width
distance in meters (m)
time in seconds (s)
frequency in radians
frequency in Hertz (Hz)
conductivity in mho per meter (mho/m)
dielectric constant in free space
relative dielectric constant
coefficient of thermal expansion
thermoelastic wave pressure
thermoelastic wave displacement
29
transmission of sound is independent of frequency and depends primarily
on the compressibility and density of the medium.
It increases with a
tissue's structural protein content, and decreases with its water
content (Leeman, 1986).
TABLE IV
lists some representative data
published in numerous journals and books.
"Attenuation"
scattering
and
refers
to
represents
the
losses
total
due
to
both
transmission
diffractive and refractive contributions.
absorption
loss,
and
including
Absorption describes the
conversion of sound into heat, which is determined by the frequency of
the sound, the viscosity of the conducting medium, and the "relaxation
time" of the medium.
Ultrasound scattering by tissue is caused by
inhomogeneity in the tissue such as variation in the local density and
elasticity.
They can be described as functions of acoustic impedance
and velocity fluctuations.
The
linear
attenuation coefficient of longitudinal ultrasound
waves in mammalian tissue, aa is expressed as a simple power of the
frequency f, such that
aa - afn
(2.4)
The constants a and n depend on tissue type, and some representative
values are listed in Table IV as reported in Leeman (1986).
(1986) also
points
out
that
attenuation,
increasing water content of the
in
general,
Leeman
drops with
tissue and rises with increasing
structural protein content, following a similar dependence to that of
the velocity.
30
TABLE IV
ACOUSTIC PROPERTIES OF BIOLOGICAL MATERIALS
(borrowed from Leeman, 1986)
Material
Attenuation/Absorption
(m_1)
Velocity
(m/s)
Water(20 °)
0.025^
1483
Blood
2.3f1-3
1560
Soft Tissue
3 - 20 f
1545-1630
Lung
lOOf1•7
2700-4100
Bone
450f°•7
650(1 MHz)
(f = frequency in MHz)
31
2.2 Microwave Absorption Pattern And Induced Thermoelastic Pressure
Waves
For a given material, microwave absorption pattern varies with
different
microwave
frequencies
and
pulse
widths,
which
in
turn
generate a different induced thermoelastic pressure wave frequency
spectrum.
of
the
A complete theoretical analysis and some experimental data
displacement
and
pressure
generated
by
microwave-induced
thermoelastic expansion in a semi-infinite medium are given in Gournay
(1966), Borth and Cain (1977), Lin (1978), and Olsen and Lin (1983) .
The
spectral
contents of
interest in this study.
these
pressure
waves
are
of particular
Simple calculation and experiments were
repeated to verify the present system and to help study the potential
of this imaging modality.
Assuming a simple model as in Fig. 1, a plane wave impinges
normally on a semi-infinite homogeneous tissue medium.
waves
thus
generated
characterized
u(x,t).
by
by
their
the
thermal
pressure,
expansion
p(x,t),
and
The acoustic
mechanism
their
can
be
displacement,
The following equations are given in Borth and Cain (1977).
„
p(x,t)=
av
_ -27X
0e
-{[sinh 27ut]U(t)
S
- [sinh 2 7 «(t-T)]U(t-T)}>
+ r "r(P.t) . 3ovlo
L
C
S
- r "rC.t) . 3«»Io
L
C
S
slnh
2-yu(t-^) 1 . U(t-^)
V
v
-1
sirll 2av(t-T-*)"| . D(t-T-*>
v
J
V
(2.5)
32
and
u(x,t)- 3 ^ [ o e ^ ^slnh_27«t _t)
S
2I>7
u(t)
+ [(t-T)-sirih 27u(t-T)] . U(t-T)}•
+ 1"
2Ir(0,t) (t-*)
vc
+
~|~21 r( 0 , t> ( t -T-*) +
L
vc
v
3aI°
2vy S
v
.(1 - cosh 2 au (t-*))l . U(t-*)
v -*
u
' (1-cosh 27u( t -T-*))l .U( t -T-*)
J
S
v -•
v
(2.6)
3aI°
2vy
The Fourier transform of p(x,t) describes the spectral contents of
these pressure waves.
F [ p < * , t ) ] - ™°">2^0 ril-cosQT)] -I
S
L CO2-
+ Uv2y2
J
(2.7)
The parameters of the above equations are explained in Table III.
The magnitudes of the Fourier transforms of these pressure waves
were computed by a computer program implemented in the Intel 86380
computer.
The data were transferred to an IBM main frame computer for
plotting by a graphic package "DISSPLA".
Figure 4 is a plot of
pressure intensities as a function of acoustic frequencies for a 2450
MHz, 2 /zs pulse.
The curve is a broad spectrum of acoustic frequency
components and the pressure is the highest at about 30 KHz.
Experiment with 2450 MHz microwave frequency and 2 fis pulse width
was set up and the induced acoustic signal was received by one of the
element of our system, which will be described in chapter three.
The
signal was amplified and filtered before feeding into a Nicolet 4094A
digital scope.
About 1 ms (1.024 ms) of signal digitized at 2 MHz was
collected and analyzed by a Nicolet fast Fourier transform program.
33
FREQUENCY SPECTRUM OF
A MICROWAVE-INDUCED ACOUSTIC SIGNAL
2450 MHz, 2.0 JJ
S
>
*
y
\
s
/
/
w
j^'o.
E- ^
/
W
Z
w
E-1
Z
/
/
/
k
y
\
s
J
V
\
\
\
i
f
\1
Iff
104
lrf
10
FREQUENCY , Hz
4 Frequency spectrum of acoustic wave induced by thermoelastic
mechanism in water irradiated with 2450 MHz, 2 y.s microwave
pulse.
34
Figure
5a,b
shows
a
typical
acoustic
wave
and
the
accompanying
frequency spectrum generated at the surface of a water tank.
transducer was 14 cm underneath the microwave source.
The
Propagation
delay of 90 ps from the triggering of the microwave pulse was expected,
taking the speed of sound in water as about 1540 m/s.
contents of the signal spread from 24 KHz to 86 KHz.
The spectral
They were limited
not only by the microwave source and transducer frequency response, the
signal was band-limited at 25 KHz and 250 KHz.
Nevertheless, it
agrees, in general, with the theoretical prediction.
Similar plots for four other microwave frequencies and various
pulse widths from the calculation are included in Appendix A.
Figure 6
summarizes the results by plotting the acoustic frequencies at maximum
intensity
as
frequencies.
a
function
of
pulse
widths
for
various
microwave
The shorter and higher microwave frequency pulses appear
to stimulate more high frequency components.
pressures occur at higher acoustic frequencies.
The maximum intensity
Therefore, a system
with acoustic signal in the diagnostic ultrasound frequency range can
be achieved by using a high microwave frequency and short pulse width
microwave source and a good quality, high frequency response transducer
as a receiver .
(b)
Fig. 5 Acoustic wave induced at the surface of the water
irradiated with 2450 MHz microwave pulses: t0 = 2
average power = 30 KV, and band-limited at 25 KHZ
250 KHz. (a) The received wave; (b) its frequency
tank
ps,
and
spectrum.
36
Acoustic Frequency and Pulse Width
Acoustic Freq., KHz
• 433 MHz
• 245(3 MHz
P 915 MHz
O 5000 MHz
A 8000 MHz
100-tAA^-
80 -
60 -
40 J
20 -
10
15
Pulse Width, microseconds
Fig. 6 Acoustic frequency at maximum intensity as a function of
pulse width
CHAPTER 3
HARDWARE SYSTEM
A block diagram of the thermoelastic imaging system is shown in
Fig. 7.
Major components of the system include a microwave pulse
generator,
a
piezoelectric
receiving
transducer
conditioning and data conversion interface, and
array,
the
intensity
of
the
thermoelastic
signal
a processing and
control computer with graphic display capability.
detect
a
The transducers
wave.
The
signal
conditioning and data conversion system conditions and translates the
signals into a set of numerically coded values in a format acceptable
to the computer.
Current signals received from
the transducer array are first
amplified and converted to voltages of useful level. The signals then
pass through a filtering section to band-limit the frequency components
to avoid aliasing and to reduce high frequency noise.
The processed
analog-to-digital converter converts these values into digital form.
The resultant digital words are transmitted to a computer data base for
storage and further processing.
In this chapter, specific design details of the hardware aspect of
the
thermoelastic
imaging
system
are
provided.
Suggestions
and
recommendations on hardware design of the system are included.
3.1 Microwave Source And Its Control Circuit
A microwave pulsed signal source ( EPCO PH40K ) designed by
Applied
Microwave
provides
microwave
37
energy for
the
thermoelastic
38
o
MUX
MUX
400 INPUTS
FROM'
TRANSDUCERS
TIMING & CONTROL
Fig. 7 A block diagram of the imaging system
A/D
39
imaging system.
The product specifications are given in Table V.
The
basic circuits of the units are shown in the block diagram of Fig. 8.
The pulse forming network (PFN) generated a pulse train which is
controlled externally to activate the driver/modulator circuit.
The
driver/modulator then amplifies the pulse from the PFN to provide a 0
to -8 kilovolts pulse for modulating a plug-in oscillator.
The power
supply provides all voltages required for operation of the equipment.
The
control
and
monitoring
circuits
provide
for
instrument as well as external oscillator voltages.
control
of
the
Front panel meters
monitor the peak voltage output, average output current, and filament
voltage to the plug-in oscillator.
A plug-in oscillator which was calibrated by the manufacturer
provides
easy
2757 MHz.
selection
of
microwave
frequency
from 487
MHz
to
With the pulse forming network operating in external pulse
mode, the pulse width, shape and period are all controlled by a system
consisting of a combination of a pulse generator (Tektronic PG 505), a
function generator (Tektronic FG 501), and a computer (Intel 86/380).
Figure 9 shows the block diagram of this external control circuit for
the pulse forming network.
Both PG 505 and FG 501 were designed to
operate in a TM 500 series power module.
The computer provides a TTL
level (Transistor Transistor Logic: 0 volt for Low level; 5 volts for
high level) signal to input PG 505, which generates a signal through
the "TRIG OUT" to the "GATE IN" of FG 501.
The FG 501 then generates a
TTL signal to the "EXTERNAL PULSE IN" of the microwave pulsed signal
source.
With this combination, a specific pulse period, duration, rise
time, fall time, amplitude, and output polarity can be adjusted easily.
TABLE V
MICROWAVE GENERATOR PRODUCT SPECIFICATION
Characteristic
Capability or Limitation
Power requirement
115 vrms, 60 Hz, 10 Amperes
Frequency range
487 MHz - 2757 MHz
Pulse width
0.4 fis - 25 us
Repetition rate
10 - 10000 pulses /second
Duty cycle
0.003
Pulse voltage output
-2 KV - -8 KV at 0.003 duty cycle
Maximum pulse current
output
12 Amperes at -8 KV
RF power output
40 KW
41
OUTPUTS: VIV E O OOT, 5YKJC OlST
_n_
(p
VAO
60
Fig. 8 A block diagram of the microwave generator
42
PULSE
GENERATOR
PG505
FUNCTION
GENERATOR
FG 501
MICROWAVE
GENERATOR
PHK 40
TRIG
OUT
GATE
IN
OUT
EXT PULSE
IN
IN
A
CONTROL
FROM
COMPUTER
9 A block diagram of the controlling circuit for the pulse forming
network
43
For all the experiments reported in this thesis, 2 microsecond pulse
width,
peak
voltage of 3 kilovolts, and peak power of about 30
Kilowatts were used unless otherwise stated.
The limitations of this
source generator include:
(1) maximum power output is 40 KW;
(2) pulse widths range from 0.4 ns to 25 /is; and
(3) frequency ranges from 487 MHz to 2757 MHz.
3.2 Acoustic Wave Detection
The acoustic waves are received by a 20 x 20 piezoelectric
transducer array specially fabricated by International Transducer Inc.
Each transducer was made of lead zirconate titanate with free-field
voltage sensitivity of -200 db at 70 KHz and is 9 mm in diameter x 2 mm
in thickness.
The array is sealed within a square metal frame that
measures 21 cm on a side.
Each transducer can be accessed directly
through a single coaxial cable.
We are interested in the peak amplitudes of the first arrival
waves.
They fall in the range of 1.5 V to 4.0 V after amplification.
Figure 10 shows a typical wave received.
It has an initial artifact,
which is believed to be caused by the triggering of the microwave
generator.
This artifact dies off in about 50 ns typically, depending
mainly on the frequency response of the transducer.
Following the
initial artifact, an ultrasound wave appears at around 96 fjts and dies
off at 150 ns.
image
sensor
stability.
Since no mechanical scanning is involved, this area
offers
reliable
measurement
and
excellent
geometric
The transducers packed together in a rather small area.
Cross talk among transducers can create serious problems.
Careful
. 10 A thermoelastic wave received by the system
45
shielding and good grounding are necessary to keep the system from
oscillation problems.
3.3 Signal Processing And Data Conversion
The acoustic waves received from the piezoelectric transducers
array must be preprocessed and digitally coded so that the computer can
further process them and display a complete image on the color monitor.
The preprocessing stages are (1) the charge amplifiers; (2) the passive
band-pass filters; and (3) the amplifiers.
For economic purpose, two
levels of multiplexers (MUX) are used to direct each channel to an
analog-to-digital converter (ADC) for digital coding.
Sample-hold
(S/H) circuits and an ADC convert the analog signals into digital code
form.
Then the digital codes are stored in the computer through an
Intel 517 parallel and serial I/O (input/output) board.
The circuits
mentioned above will all be described in this section.
3.3.1 Charge Amplifiers
Piezoelectric materials have high but finite resistance.
As a
result, the charge generated by an applied force leaks through the
leakage resistor, causing the potential difference eventually to reduce
to zero.
The attachment of a voltage amplifier to the output of the
transducer will degrade the transducer characteristics if the input
impedance
of
the voltage
magnitude higher
amplifier is not at least an order of
than that of the piezoelectric
transducer.
The
resistance of a piezoelectric transducer is typically on the order of
100 GO .
The usual input resistance becomes important, causing charge
to be lost and thereby decreasing the output voltage.
46
In order to improve the dynamic response of the transducer at low
and medium frequencies, the output of the piezoelectric transducer is
fed directly into the negative input of a charge amplifier (Cobbold,
1974).
Figure 11
circuit.
shows the schematic diagram of a charge amplifier
There is a virtual ground at the negative input to the
amplifier.
Theoretically, current generated by the transducer all
flows into the feedback capacitor and no current flows through the
input
resistor
or
the
input
capacitor.
However,
the
operational
amplifier (op-amp) bias current causes the charge amplifier to drift
slowly with time.
Therefore, the large feedback resistor Rf is added
to the system to avoid saturation problems.
Operational
amplifiers
Semiconductor LF356.
high
input
in
these
circuits
are
National
They were chosen for their low power consumption,
impedance,
characteristics.
used
high
frequency
response
and
low
noise
The product specifications are given in National
Semiconductor (1982).
The charge amplifier also behaves as a low pass filter with cut­
off
frequency
equal
to
1/2 RfC.
The
highest
frequency
of
the
transducer array within its linear range is about 250 KHz.
3.3.2 Band-pass Filters
The equivalent circuit for a piezoelectric transducer at high
frequencies is more complex due to its mechanical resonance.
The
dynamic response of the transducer at high frequencies is not a simple
linear function.
In order to avoid aliasing and high-frequency noise,
a filtering section is introduced to band-limit the signal.
Previous
experiments have shown that the frequency range of the transducer is
Fig. 11 A charge amplifier
48
linear - approximately between 25 KHz to 250 KHz (Olsen and Lin, 1983).
A 4th order band-pass filter with cutoff frequencies of 25 KHz and 250
KHz was selected.
The overall bandwidth of the filter is 225 KHz,
while the percentage bandwidth is 285 percent.
Since the percentage
bandwidth is more than 80 to 100 percent, an overlapping high pass and
low pass cascaded filter is desirable.
advantages
over
the
traditional
passive
Active filters have many
filters.
However,
the
frequency response attainable in most commercial active filters is
limited to 50 KHz.
For this application, passive filters were chosen
(Fig. 12).
Passive filter has a drawback of bad isolation between stages and
its current gain is not always provided.
isolation,
dual
N-channel
JFET
2N5912
In order to improve stage
designed
for
wideband
differential amplifiers from Siliconix Incorporated were chosen as
source-followers (Siliconix, 1982) to buffer between stages and provide
proper current gain to the filter circuit (see Fig. 68 in Appendix B).
A source-follower circuit needs to exhibit high input impedance and low
output impedance.
The real part of the output impedance is the
reciprocal of common-source forward transconductance, which is, for
this case, about 200 Q up to about 600 MHz.
The input capacitance is
approximately 1.5 pF (maximum) and is independent of frequency and load
when the load is larger than the output resistance.
The frequency
response is dependent mainly on the output impedance of the previous
stage.
The low-frequency voltage gain of this particular circuit is
calculated to be about 0.9.
Thus the overall voltage gain of a
filtering section is less than 0.78.
49
out
in O
f L = 250 KHz
fR =
25 KHz
Passive Bandpass Filter
Fig. 12 A cascaded low-pass, high-pass passive filter
50
The performance of this filtering circuit has been tested by
feeding a 0.25 V sine wave at the input of the first stage inverted
amplifier (described in the next section) and measuring the output of
the second stage inverted amplifier.
the magnitude of the
Figure 13 shows the Bode plot of
transfer gain of a filtering and amplifier
section.
The 3 dB points are at 25 KHz and 170 KHz and are rolling
down
6
at
dB/decade.
The
filter
meets
the
cut-off
conditions, however, it is not falling off sharp enough.
seen that great improvement is needed.
It is clearly
Figure 14a and 14b attempts to
show the phase characteristics of the filter.
detected and its effect is negligible.
frequency
A minimum phase shift is
A completed schematic diagram
of this band-pass passive filter can be found in Appendix B.
3.3.3 Amplifiers
Two stages of inverted amplifiers were implemented for proper
amplification and impedance matching ( Fig. 15).
of this circuit is -Rc/R£.
and Rc respectively.
The transfer function
For simplicity, we used 1 KQ and 10 KQ for
Therefore, each stage has a gain of ten.
The
first stage is placed between the first level MUX (which will be
discussed later) and the filter, and the second stage is after the
filter
and
before
the
S/H.
Texas Instruments TI082
amplifiers were chosen for this application.
dual
BIFET
This BIFET amplifier has
high input impedance, low noise, low power consumption, small in size,
high frequency response and excellent stability.
The detail product
specification can be found in Texas Instruments (1985).
this point has a typical value of 2-4 V.
Signal up to
51
BANDPASS FILTER FREQUENCY RESPONSE
*-H _
1-)
r—l
\
/
V
s
/
/
\
"~71
....
\\
/
/
\
/
/
/
\
r—i
10
10
10°
FREQUENCY , Hz
Fig. 13 A Bode plot of the magnitude transfer gain of a filtering
and amplifier section
(b)
Fig. 14 Phase characteristic of the filter, (a) input; (b) output
53
1000J1.
Fig. 15 An inverted amplifier
54
3.3.4 Multiplexers
There are two levels of a total of 27 multiplexers.
Analog
Devices AD7506 monolithic CMOS 16-channel analog multiplexers were
chosen.
The first level MUX selects 25 out of the 400 input channels
to 25 amplify, filtering, and S/H circuits. The second level MUX
switches serially one of these 25 signal conditioning sections to the
analog-to digital converter for digital coding.
Figure 16
is the
functional diagram of an AD7506.
These CMOS switches have extremely
low quiescent power dissipation.
Their ohmic resistance between the
output and an addressed input is low (300 ohm); in the off condition,
leakage is quite small, both across the gate and to the drive and
supply circuits.
The switching of each MUX is controlled by four
address lines and an "ENABLE" line.
TTL/DTL, as well as CMOS logic.
1-2
KD
are
compatibility.
recommended
Data
by
Those control lines respond to
However, pull-up resistors, typically
the
acquisition
manufacturer
routines
to
ensure
explained
in
TTL/DTL
the
next
chapter, overlook all the events taken place for data acquisition.
They supply proper addresses to the first level multiplexers, while the
second level multiplexers are controlled directly by a counter TI74193.
With this hybrid parallel and serial combination design, the cost was
reduced tremendously.
3.3.5 Data Conversion
The
CMOS
microprocessor-compatible
analog-to-digital
converter
AD7574 by Analog Devices uses the successive-approximation technique to
quantize the wave signal into the appropriate 8-bit digital code in
15 us (see Fig. 17 for the circuit diagram of the ADC).
Twenty-five
en a3 a2 a1 ao
G> Q
G> Q Q
r
vdd q
(+15v)
-i
t t l /nTI T0 CM0S LEVEL
|
TRANSLATOR
gnd G—'
|
VSS
C-15V)®~l
DECODER/DRIVER
r
i
6 6
OUT SL
s18
16 Functional block diagram of AD7506
Q +5V
I(j>\C
rvVvS
AP 7574
(TOP VIEW)
nAAA-*-^
g7 ZV.
-b
IczSv
-HgSVf
te^F
AP 584
3
vc,
^ K,
a A a S
3i0c
I\
5l6MAtlMPUT
0
-|0V-fcf/^°
126 5k~
ANALOG
6£OUND
<SA|M Tf2-iM
AM AL-OQ
r5UPPi.Y eetUSH
GROUND lWTEer/£
D16ITIAU
5UPPLY |2£TUai"
Fig. 17 A schematic diagram of an analog-to-digital converter
57
Datel Intersil
SHM-IC-1 . sample-and
holds, which
have an aperture
uncertainty of 5 ns, hold samples constant during conversion period.
With this combination, the maximum input signal acquisition frequency
of 250 KHz can be achieved (Analog Devices, 8,pp. 14-20,1982).
The AD7574 operates from a +5 volts power supply and has an
internal clock, an on-board comparator, and interface logic.
The +10
volts and -10 volts references are provided by a pin programmable
precision voltage reference AD584 (from Analog Device) and an inverter
circuit.
The analog resolution, which is a measure of the nominal
analog change required for a 1-bit change in the ADC's digital output,
is 0.078 volts for the ADC operating in bipolar mode.
The AD7574 is operated as a static Random-Access Memory (RAM).
The RAM interface mode uses distinctly different commands to start
conversion.
The timing diagram is shown in Fig. 18.
A convert start
is initiated by executing a memory WRITE instruction to the AD7574,
which causes CS (pin 16) to go low.
the conversion is completed.
The BUSY" (pin 14) goes low until
Once BUSY" goes back to high, a data read
is performed by executing a memory READ instruction to the AD7574
address location.
read
before
a
When a conversion is finished the fresh data must be
new
conversion
manufacturer recommends
can
be
started.
that a memory READ
to
Therefore,
the
the AD7574 address
location be executed and subsequently the data upon initialization
after power up be ignored.
The internal synchronous clock oscillator
starts
start
once
the
convert
command
oscillating when conversion is complete.
is
Rclk
=
received
and
ceases
^25 Kfl, and Ccik = 100
pF are required to ensure a conversion time of 15
us.
MIC20PR.0CES50EOPERATION
MEMORY WRITE
TO AD7574 ADDRESS
CS CPIN 16)
NOP OR OTHER INSTRUCTIONS
UNTIL BUSY IS H16U
v
itencs-
^esp"V
ED(PIKJ15)
^coMveer
t^WCS *
Su
j*"
-f2£5eT
u
\
*-*3505 - --teAD
-*CPBD»
DB7 TO OB0
(PIN6 13?
TO AP7574 APP££S6
jr
X.
BUSY CPlW 14)
NEMOSY WEiTE
MEMORY WRITE
TO AP7574 APPRESS
HIGH Z
-•tgHO
v DATA
A
Fig. 18 Timing diagram for the analog-to-digital converter
_A_
HIGH Z
59
The SHM-IC-1 (Fig. 19) is a self-contained device that requires
only an external hold capacitor.
The value is chosen by the user to
achieve the desired speed and required accuracy.
is used.
This gives an acquisition time of 4
A 0.001 fiF capacitor
us (10 v to 0.1 %) and
hold mode voltage droop of 50 mv/sec maximum.
The hold capacitor
should be of a good quality with a very high insulation resistance and
low dielectric absorption.
A polystyrene type capacitor is highly
recommended by the manufacturer (Datel Intersil, 1983/84).
The S/H's are initialize to be in the sample mode when data
acquisition has just begun.
After about 80 ^s delay from sending out a
microwave pulse, a series of one shots configured as in Fig. 69 in
Appendix B detects the arrival of an acoustic wave which in turn puts
the S/H's into hold mode.
This 80 fis delay can be adjusted through a
potentiometer on the front panel of the system.
This delay period
depends on the relative position of the applicator and the transducer
array (dre^).
Assuming that acoustic wave travels 1500 m/s in water,
the delay period (T^elay)
^delay
^rel
m
can
be calculated by the following equation:
/1500 m/s
(3.1)
f
With this delay time, the control circuit will eliminate false wave
arrival information.
3.4 Controlling Unit And Its Interface
3.4.1 Microcomputer
The Intel 86380 microcomputer incorporates the Intel 8086 as the
central processing unit (CPU).
It runs at 8 MHz under the operating
-I5VDC +I5VDC
I
!
•o OUTPUT
CASE
-ISV
Fig. 19 A schematic diagram of a sample-and-hold
61
system iRMX86 release 5.
It has 896 Kbyte Random-Access Memory (RAM),
a 20 Mbyte Winchester drive and a 1 Mbyte 8 inch floppy drive.
The 8086 is Intel's general purpose 16-bit microprocessor.
It has
a 20-bit wide address bus, providing capability of addressing a full
megabyte of memory.
The upper 4 bits of address share their pins with
some status signals, and the lower 16 bits are shared pins with data.
It consists
of fourteen
16-bit
internal
registers.
Its "number
crunching" capability relies heavily on its co-processor 8087 numeric
data processor (NDP).
The 8087 NDP effectively adds eight 80-bit
floating point registers to the 8086 register set.
It uses its own
instruction queue to monitor the 8086 instruction stream, executing
only those instructions intended for the 8086 CPU.
The 8087 NDP
instructions include a full set of arithmetic functions as well as a
powerful core of exponential, logarithmic, and trigonometric functions.
This
capability
is
essential
for
the
current
image
processing
application.
The Intel iRMX 86 operating system is an easy-to-use, real time
and
multitasking
software
system.
It
is
designed
to
provide
a
structured and efficient environment for many time and performance
critical applications such as controlling devices, process control, and
factory
automation.
interpreters
with
The
Intel's
system
supports
Universal
language
Development
compilers
Interface
and
(UDI).
Currently three high level languages, FORTRAN, PL/M, and Pascal, and
iAPX86 Macro Assembler are installed.
environments:
debugger).
static (iSBC
957
B
It also supports two debugging
monitor) and
dynamic (iRMX
86
The static debugging can debug one task at a time while the
62
dynamic debugger provides evaluation and testing of a real-time system.
The iSBC 975 B monitor is a stand alone monitor for static debugging.
All debugging instructions necessary to view and modify register and
memory contents are accessible from a terminal connected directly to
the system.
The iRMX 86 debugger runs as a part of an iRMX 86
application.
It allows programmers to stop and inspect one task while
the rest of the system operates normally.
The major drawback of this system lies in its software design
rather than its hardware design.
They include an inefficient operating
system, which takes up more than half of the on board memory, a poor
file management feature, and a lack of commercial software support.
These factors all contribute to long development time.
3.4.2 Input/output Expansion Board
The iSBC 517 combination I/O Expansion Board provides the I/O
interface to the data acquisition circuit of the thermoelastic tissue
imaging system.
It was designed to interface directly to the Intel
86380 microcomputer to expand system interrupt levels and serial and
parallel I/O lines.
The parallel I/O expansion feature is utilized in
the thermoelastic imaging system.
Therefore only this feature will be
discussed in this section.
The parallel I/O expansion feature incorporates two Intel 8255
Programmable Peripheral Interfaces.
Together, these devices provide 48
I/O lines, which can be independently programmed by the system software
with the three ports A, B, and C.
These 48 lines can be implemented in
various combinations of unidirectional input/output and bidirectional
63
ports, by which the software can program to operate in any of these
three modes.
Sixteen of the 48 I/O lines (port A's) have bidirectional driver
(Intel 8226) and termination networks permanently installed.
One of
these port A's is connected to the data bus of the ADC AD7574 and is
programmed
as
input
port.
Sockets
are
provided
for
installing
interchangeable quad line drivers and terminators for the other 32 I/O
lines.
This provision allows the user
to select sink currents,
polarities and other characteristics appropriate to the application.
Currently, two (SBC901) terminating pack and two NAND gates (7400) are
installed in ports B and C of the same 8255 respectively.
Port B is
configured as output port, while port C is an input port.
Figure 20
shows the function of each I/O line.
I/O's.
The are configured as I/O mapped
TABLE VI gives the I/O base address assignment.
Control
routines, which will be described in the next chapter, initialize and
access the memory and I/O facilities on the iSBC 517. See iSBC 517
Combination I/O Expansion Board Hardware Reference Manual for detail
description.
3.4.3 Other Circuits
A combination of a counter 74193 and a D flip-flop 74175 is used
to control the counting of the second level MUX.
Appendix B shows a schematic diagram of this circuit.
Figure 69 in
The count-up
line is triggered at the rising edge of BUSY" of the ADC AD7574.
The
data outputs of this counter provide the proper addresses to the
multiplexers.
The flip-flop is set when the carry bit of the counter
64
0
1
2
PORT 3
A
4
5
6
7
PORT
B
<
<
<
<
<
<
<
<
0
1
2
3
4
5
6
7
0
1
2
PORT 3
C
4
5
6
7
FROM A/D
CONVERTER
>Ao
>Ai MUX (1st Level)
>A2 CONTROL LINES
>A3
>
CLEAR THE COUNTER
>
CLEAR THE CARRY BIT
>
PULSE OUT
>
NC
<
<
<
<
<
<
<
<
Fig. '20 Functional diagram of I/O lines
BXJSY FROM A/D CONVERTER
NC
START FROM FRONT PANEL
START CONVERSION
Nc
NC
NC
NC
65
TABLE VI
INPUT/OUTPUT BASE ADDRESS ASSIGNMENT
ADDRESS
PORT 1
PORT 2
PORT 3
CONTROL
REGISTER
B4
B5
B6
B7
66
turns high, which in turn switches on the second multiplexer and off
the first one.
3.5 Display
3.5.1 Color Monitor
Color II monitor is a medium resolution CRT (cathode ray tube)
designed by Amdek Inc. The RGB (Red, Green,and
Blue)inputs permit up
to 16 computet controlled color graphics and 80 x 24 color text
display.
Line resolution is 560(H) x 240(V).
All the control lines
and color inputs accept positive TTL signals and are easily interfaced
to a graphic controller described in the next section.
3.5.2 Graphic Controller
The RGB-Graph is a multibus compatible color video boards from
Matrox
Electronics
Systems Ltd.
It has 16-bit 512 x 512 pixel
resolution and writing speed of 800 ns/pixel.
The board contains
advanced video features such as hardware zoom, scroll, shift, pan,
clipping, overlay and video masking.
It is mainly used as a display
buffer (128 Kbyte on board memory) and controller for the color II
monitor.
It is presently configured as a I/O mapped I/O device and
connected to the Intel 86380 expansion slot directly.
The resolution of both
the graphic controller
monitor is sufficient for the application.
and the color
However, the limited 16
display colors make images look more discrete than they actually are.
CHAPTER 4
SOFTWARE
The ACQ4 is a software package consisting of Pascal-callable
subroutines
and
utility
programs
thermoelastic imaging system prototype.
for
the
microwave-induced
All routines were written in
Intel PASC86 Pascal language or ASM86 assembler language.
The ACQ4 is
intended to be a user friendly and easily expandable software. Menus
will be displayed on the terminal and the program will prompt the user
for specific information.
It was designed specifically to allow data
acquisition from 20 x 20 piezoelectric transducer array, interactive
manipulation of the
acquired image data, and display pseudo-color
images on an Amdek Color II monitor through a RGB-graph image board
from Matrox.
However, most of the routines can handle different sizes
of matrix, except those directly related to the display system.
Simple
modification will enable this software package to adapt to other image
processing applications.
The objective of this chapter is to describe the uses and the
implementation of these routines to an Intel 86380 microcomputer as
part
of
the microwave-induced
thermoelastic tissue imaging system
prototype at the Bioengineering Department, University of Illinois at
Chicago.
Most of the mathematical derivation of these image processing
techniques have been presented in many digital image processing books,
therefore it will not be repeated here. For a more complete discussion
of the techniques and programming methods, see references Rosenfeld and
67
68
Kak (1982), Hall (1979), Duda and Hart (1973) and Appendix C for the
source listing.
Figure 21 represents an overall structure of ACQ4 software package
and serves as a guide line to the organization of this chapter.
This
chapter is divided into four major sections: (1) data acquisition, (2)
subtraction and normalization, (3) image processing, and (4) display.
Examples of acquiring and processing of images of a glass tube filled
with
tissue
equivalent
material
are
included
to
demonstrate
the
performance of this software package.
4.1 Data Acquisition
The main function of the data acquisition routine, refer as DAR
throughout the rest of the text, is to coordinate all the events taken
place while the computer is acquiring raw data of an image.
selected from the first menu of the main program.
counter,
flip-flops
and I/O (input/output) ports
operator that data acquisition may begin.
DAR can be
It initializes the
and signals the
Once the phantom is properly
placed, the microwave source is tuned to the desired power level, and
the operator is ready, the "START" button is pressed to prompt the
computer to begin the data acquisition.
DAR
signals
the
function
generator
to
send
a
pulse
to
the
microwave pulse generator which in turn delivers a 2450 MHz microwave
pulse to the water tank through an applicator.
Thermoelastic wave thus
generated, which has been explained in chapter two, propagates towards
the transducer array.
After a delay time set by the operator through
the potentiometer on the front panel, DAR checks the arrival of this
wave at one of the transducers. As soon as DAR receives the wave
69
ACQ4
IMAGE
PROCESSING
DATA
ACQUISITION
SUBTRACTION /
NORMALIZATION
ENHANCEMENT
LINEAR IMAGE
MANIPULATION
SMOOTHING
FILTERING
EDGE
ENHANCEMENT
FILTERING
DISPLAY
AVERAGE
CHECK
THRESHOLDING
NON-LINEAR IMAGE
MANIPULATION
LINEAR
GREY SCALE
TRANSFORM­
ATION
INTERACTIVE
MODE
Fig. 21 An overall structure of ACQ4
PIECEWISE
LINEAR
GREY SCALE
TRANSFORM­
ATION
HISTOGRAM
EQUALIZA­
TION
70
arrival information, it puts all 25 S/H's into the hold-mode.
directs
It also
the multiplexers which sequentially switch between the 25
channels, and starts the ADC conversion.
digital data into memory.
very crucial.
Finally, it stores the
The time sequence of each of these events is
DAR monitors everything to make sure that multiplexers
are not switching while the ADC is converting; another ADC conversion
will not start until previous converted digital data have been stored
into memory and multiplexers have switched to a new channel. The above
process has to be repeated 16 times for each image.
DAR then returns
to the main program when one complete image has been taken.
The
program prompts the operator to store the image into a data file or to
retake the picture.
4.2 Subtraction and Normalization
The background image can be considered as the source function (see
Fig. 22).
the
It is noted that an acoustic wave front propagating towards
transducer
is
not
evenly
distributed
throughout
transducer array, rather it radiates out from the center.
the
entire
Area at the
center of the transducer array receives stronger signal than on the
sides.
As discussed in the previous chapters, this acoustic source is
a function of microwave frequency, power density, pulse width and the
I
material irradiated.
The first three parameters can be selected from
the pulse generator and its controlling circuit.
In traditional ultrasound imaging, subtraction of a data set with
no object present from those with objects present is used to remove the
variation
of
simplicity,
it
the
was
source
function (Johnson
assumed
that
the
et
image with
al.
1979).
For
phantoms present
Fig. 22 A "BACKGROUND" image
72
("OBJECT"
image) represents the source function attenuated by the
phantoms.
Subtraction of this "OBJECT" image from the "BACKGROUND"
image thus
equals
the attenuation profile of the phantoms.
For
example, Fig. 23 is an image of a 1.8 cm test tube filled with muscleequivalent
material
and
Fig.
24 is
its subtraction image.
The
attenuation profile of this phantom can easily be seen from this noisy
subtraction image.
The noise may come from different sources such as
electronic noise, thermal noise, cross talk or channel inconsistency.
Electronic noise, thermal noise and cross talk may be removed by
convolution filtering, which will be discussed in the next section
(Johnson et al. 1979).
Discrepancies in discrete components design
and transducer element beam pattern variation may contribute mainly to
the channel inconsistency problem.
to these problems.
Normalization may offer a solution
Two different normalization techniques have been
explored:
(1) B(xi,yi) - 0(xi,yi) .
for
.=
lf
2Q
B(xi,yi)
B(xi,yi) - 0(xi,yi)
(4.1)
_
(2) Ln I"
; for i=l, 20
L
B(Xi,yi)
J
(4.2)
B = background image, 0= image with object.
For a particular pixel i and processing channel j,
let
where
B(xi,yi) = l0ie"abXiGciGfjGaj!
(4.3)
°(xi>yi) =
(4.4)
Iq i
ajj
at
x^
xtj
IOi e a
" b(xi"xt)e"atxtGciGfjGaj;
= initial response at the water surface of the ith
transducer,
= attenuation coefficient of water (background),
= attenuation coefficient of material inside the tube,
= total distance travelled,
= distance travelled inside the tube,
73
Fig. 23 An "OBJECT" image: a 1.8 cm glass tube filled with muscle
equivalent material
75
Gci
Gfj
. = gain of the ith charge amplifier,
= gain of the jth filtering section,
= gain of the jth amplifier section.
Gaj
A simple subtraction of 0(xi,yi) from B(xj[,yi) becomes
B(xi,yi) - 0(xiiyi) = IOiGciGfjGaj(e-%x -
Q
x x
a x
e- b( " ti)e- t ti)
(4.5)
For x » xt, equation (4.5) becomes
B(xi.yi)
" 0(*i,yi) = I0iGciGfjGaje-abx (1 - e-atxti)
(4.6)
Iq i , Gc£, Gfj, Gaj vary from channel to channel, it
Since
contributes some errors to the image.
With
the
first normalization
technique, the normalized pixel
becomes
B(xi»yi)
" °(xi.yi)
-a^x*
atxti
= 1 - e~
B(xi)yi)
Theoretically,
the
(4.7)
non-uniform
illumination
due
to
unequal
wave
propagation distances, and non-uniform gains for different parts of the
processing
channels
have
been
cancelled
by
this
normalization
technique.
The pixel value depends only on the attenuation coefficient
of the material inside the tube and the distance travelled inside the
tube.
With the second normalization technique, the normalized pixel
becomes
In
r
1-
"(xt.yj) - CKxj.yj) _ j ,
B(Xi,yi)
J
(4.8)
It is also independent of unequal wave propagation distances and non­
uniform gains for different parts of the processing channels.
The
pixel value is linearly proportional to the attenuation coefficient of
76
the material and the distance travelled inside the tube.
Figure 25
plots the first normalization function as a function of the second
normalization function.
When atxt is small, the two functions bears a
linear relationship with each other.
It implies that data fall into
this category, the two normalization methods will be equally effective.
When atxt increases to infinity, the function 1 - e"atxt approaches 1
and is independent of atxt.
In that case, the second normalization
will result in a better dynamic range of values.
For our application,
atxt has been kept small and it was found that both techniques are
equally effective.
Figure 26 and 27 show the processed images.
4.3 Image Processing
One of the most important goals of designing image processing
routines (IPR), applied to these original images, is to be able to
identify the phantom from the processed images.
In general, the
concept of image processing can be defined as processes or operations
which modify an image or groups of images to enhance the visibility of
useful information while suppressing non-useful information, or so
called "noise".
Those operators can be classified into two main
categories: (I) linear image manipulation (or convolution filter);
(2) non-linear image transformation (or grey scale transformation).
A
significant number of image processing techniques have been developed
and evaluated for specific applications.
These applications include,
for example, image-smoothing techniques (Mastin, 1985; Chin and Yeh,
1983; Rosenfeld and Kak, 1982; Huang, 1980; Panda, 1978a; Panda,
1978b), edge enhancement/detection (Peli and Malah, 1982; Davis, 1975),
grey scale manipulation (Rosenfeld and Kak, 1982; Frei, 1977; Hummel,
77
First Method
0.4-
•
0
|
0
2
|
|
]
|
|
4
6
8
10
12
14
' Second Method
Fig. 25 A plot of the first normalization function as a function
of the second normalization function
swwsanias
Fig. 26 A normalized image of Fig. 24 - the first normalization
technique
79
Fig. 27 A normalized image of Fig. 24 - the second normalization
technique
80
1975; Duda and Hart, 1973), and thresholding (Weszka, 1978; Panda,
1978b).
Several of these techniques have been implemented in the
system and will be briefly discussed in the following subsections.
4.3.1 Linear Image Manipulation
Linear image processing can be described as a two-dimensional
convolution of the
image with a two dimensional filter function.
Different filter functions selectively enhance or remove information of
different spatial
frequencies.
A
total
of 27
functions have been implemented in the system.
predefined filter
They consist of various
kinds of low pass filters, high pass filters, and edge enhancement
filters.
When a certain type of filter is chosen, a brief description
of all the predefined filters of this specific type will be displayed
on the screen and the operator is then prompt to select one of them.
All
the
predefined
filter
functions
are
stored
in
a
data
file
"IPl.data", and will be read into the program when filtering routines
are being called.
They were chosen as the basic filtering functions
and can be easily modified if a different application arises.
Filters,
other than those predefined filter functions, can also be used when
interactive mode is entered.
Images collected are always subject to noise and interference from
different sources
sensor noise.
appears
as
including
electronic
noise, channel errors, and
Image noise arising from the above sources usually
discrete
isolated
pixel variations
that are spatially
decorrelated and generally has a higher frequency spectrum than the
normal image components.
be
effective
for
noise
Hence, simple low-pass spatial filtering can
smoothing.
Six
smoothing
filters
are
implemented.
Three of them weight all their neighbors equally while
the rest put different weightings for the sampling point and its
immediate neighbors.
Figure 28 shows all these six filter functions.
The first one replaces each pixel of the original image by the sum of
its 8 neighbors and itself.
The second one is the same as the first
one except that the array is normalized to unit weighting so that the
low pass filtering process does not introduce a brightness bias in the
processed image.
neighbors.
The third one considers only its four immediate
The fourth one weighs itself two times as much as each of
its 8 neighbors.
The fifth one weighs itself one fourth of the total
weight, its 4 immediate neighbors each with one eighth of the total
weight and one sixteenth for the four corners of its 8 neighbors.
With
this weighting, it is assumed that the degree of influence by the
neighboring pixels decreases to the relative distance from the pixel.
The last one is a smoothing operator formed by repeated application of
the operator [1,1] two times in both horizontal and vertical directions
and normalized to unity.
It is noted that multiple applications of
this operator produce better and better binomial approximations to a
normal smoothing filter.
Figure 29
is the processed image of Fig. 26a
by the last operator.
Images
with
accentuated
pleasing to our sight.
edges
may
sometimes
be
subjectively
Two major types of edge operators were
implemented in the system: (1) directional; (2) non-directional.
The
most common and historically earliest edge operator is the gradient.
It is an approximation to the first partial derivatives.
It can be
designed to be sensitive to edges pointing towards certain direction.
NUMBER
3X3 OPERATORS
1
1.0 1.0 1.0
1.0 1.0 1.0
1.0 1.0 1.0
2
1/9
1/9
1/9
3
4
1/9
1/9
1/9
1/9
1/9
1/9
0.0 1.0
1.0 1.0
0.0 1.0
0.1 0.1
0.0
1.0
0.0
0.1
0.1 0.2 0.1
0.1 0.1 0.1
5
1/16 1/8 1/16
1/8 1/4 1/8
1/16 1/8 1/16
6
1/16 2/16 1/16
2/16 4/16 2/16
1/16 2/16 1/16
Fig. 28 Operators for smoothing filters.
(b)
1/16 2/16 1/16
2/16 4/16 2/16
1/16 2/16 1/16
(a)
29 A filtered image of Fig. 26
(a) the operator;
(b) the processed image.
84
Line segment edge enhancement enhances lines going in a particular
direction.
The second type is the Laplacian edge detection operator.
It is an approximation to the mathematical Laplacian:
S2f/Sx2 + S2f/Sy2
(4.9)
The Laplacian does not give directional information and tends to
enhance noise in the image. It imitates the lateral inhibition process
of human eyes.
It simply subtracts its eight
without weighting from the original pixel.
neighbors with or
Figure 30 lists all the
edge enhancement filter functions implemented in this system.
For example, it is known that the output of an operator of the
form shown in Fig. 31a is zero when it is applied to a region of
constant, or even to a linear ramp.
The output will be non-zero when
it is placed over an edge between two regions of uniform density.
Applying
this
filter
to
sharpened (see Fig. 31b).
the image shown in Fig. 31a, edges
are
However, it generates more noises as well.
4.3.2 Non-linear Imape Processing
A number of algorithms have been developed for modifying the grey
scale of digital image (Pratt, 1978).
Grey scale transformation is a
simple but powerful class of enhancement operations.
They are often
used to increase the contrast of the display image by stretching,
compressing, or shifting
the grey scales.
As a result of those
manipulations, the perception of details within the
image can be
improved.
The transformation can be expressed as a mapping from a given grey
scale into a transformed grey scale.
linear,
quadratic,
logarithmic
or
This transformation can be
completely
arbitrary.
Three
85
CATEGORIES
Gradientdirectional
3X3 OPERATORS
1.0 1.0
1.0 -2.0
1.0
1.0
-1.0 -1.0 -1.0
1.0 1.0
-1.0 -2.0
-1.0 -1.0
N
Line segment
1.0
1.0
1.0
NE
-1.0 -1.0
1.0
-1.0 -2.0
1.0 1.0
1.0
1.0
-1.0 -1.0 -1.0
1.0 -2.0
1.0 1.0
1.0
1.0
S
1.0 1.0 -1.0
1.0 -2.0 -1.0
1.0 1.0 -1.0
1.0 1.0 1.0
1.0 -2.0 -1.0
1.0 -1.0 -1.0
w
NW
2.0 -1.0
2.0 -1.0
2.0 -1.0
VERTICAL
1.0
1.0
1.0
E
SE
-1.0
-1.0
-1.0
-1.0 1.0
-1.0 -2.0
-1.0 1.0
1.0 -1.0 -1.0
1.0 -2.0 -1.0
1.0
1.0
1.0
sw
-1.0 -1.0 -1.0
-1.0 -1.0 2.0
-1.0 2.0 -1.0
2.0 -1.0 -1.0
HORIZONTAL
L-R DIAGONAL
-1.0 -1.0 -1.0
2.0
2.0
2.0
2.0 -1.0 -1.0
-1.0 2.0 -1.0
-1.0 -1.0 2.0
R-L DIAGONAL
Laplacian
-1.0 -1.0 -1.0
-1.0 8.0 -1.0
-1.0 -1.0 -1.0
0.0 -1.0 0.0
-1.0 4.0 -1.0
0.0 -1.0 0.0
1.0 -2.0 1.0
-2.0 4.0 -2.0
1.0 -2.0 1.0
others
-1.0 -1.0 -1.0
-1.0 9.0 -1.0
-1.0 -1.0 -1.0
0.0 -1.0 0.0
-1.0 5.0 -1.0
0.0 -1.0 0.0
1.0 -2.0 1.0
-2.0 5.0 -2.0
1.0 -2.0 1.0
0.0
-1.0
0.0
1.0
0.0
0.0
0.0 -1.0
0.0 1.0
0.0
0.0
-1.0
0.0
0.0
1.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Fig. 30 Operators of edge enhancement filters
(b)
-1.0 -1.0 -1.0
-1.0
8.0 -1.0
-1.0 -1.0 -1.0
(a)
31 A filtered image of Fig. 29
(a) the operator;
(b) the processed image.
87
possibilities of scaling the output image back into the domain of
values occupied by the original image have been implemented.
shows the transfer functions of
two common techniques.
Figure 32
By the first
one, the processed image is linearly mapped over its entire range.
The
transfer function can be described as the following equation:
f(gi) =
where
gj_
a
b
f(gi)
a'
b'
b
"a
b-a
x (gj[ - a) + a'
(4.10)
= the current grey level;
= the lowest grey level, usually equals 1;
= the highest grey level;
=
new transformed grey level;
= the lowest grey level of the new transformed scale;
= the highest grey level of the new transformed scale.
If a raw image falls only on a portion of the grey scale available, the
contrast of the
image can be greatly improved when the image is
transformed to the entire range linearly. For the second technique, the
processed
ranges.
image
is
piecewise
linearly
mapped
over
three separate
Each range has its own transfer function which allows specific
range of grey scales of the original image to be selectively stretched
or compressed.
For example only a portion of an image is of interest,
one can selectively stretch that range of grey scale to a larger range
of grey scales and suppress the rest.
It will result in very fine
discriminations between grey levels in the interested region and loss
of resolution in the other regions of less interest.
The third
technique rescales the original image so that the histogram of the
enhanced image is forced to be uniform and the process is also known as
histogram equalization.
Many reports (Hummel, 1975; Frei, 1977) on
histogram modification techniques have been published.
Some of them
have explored the use of these histogram modification procedures that
GREY SCALE TRANSFORMATION
NEW |
I
I
B'|
I
I
I
*
A' j*
A
*
*
*
,
B
OLD
(a)
GREY SCALE TRANSFORMATION
NEW |
D' j
C' j
I
I
B' j
I
A' |*
A
*
*
*
*•
* •
* .
.
.
. .
B C
.
D
OLD
(b)
Fig. 32 Grey scale modification methods,
(a) linear;
(b) piece-wise linear.
89
produce
enhanced
histograms.
images
with
hyperbolic
or
exponential
shaped
Figure 33 contains examples of contrast enhancement of a
tissue phantom image processed by histogram equalization.
4.3.3 Thresholding
Thresholding is often used to segment the image into regions which
can subsequently be analyzed separately.
for
a
survey
thermoelastic
of
threshold
tissue
selection.
images,
background, are considered.
two
types
See reference Weszka (1975)
For
of
application
regions,
to
the
object
and
Threshold selection in its simplest form
involves choosing a gray level t such that all grey level greater than
t are mapped into the "object" label, and all other grey levels are
mapped into the "background" level.
Since this selection depends on
the grey level of each pixel alone, the threshold is often called
global threshold.
Grey level histogram and intensity difference between two neighbor
pixels are two useful tools implemented in the system for aiding the
operator to choose a threshold.
When a threshold help menu is chosen,
the program will automatically compute a histogram of the given image
and display on the screen.
Then the program will prompt the operator
to enter either a row or column of the image for neighbor pixel
intensity difference computation.
From this intensity difference, one
may locate where edges are and enter the edge coordinates.
From these
information, the program will compute and suggest a threshold value.
It is found that the histogram technique is extremely useful to find a
threshold.
Fig. 33 A processed image of Fig. 31 by histogram equalization
91
4.4 Display
The display algorithm was designed to display a 20 x 20 image on a
color monitor.
Each image is interpolated bilinearly into a 60 x 60
matrix before transferring the data to the color monitor controller.
Bilinear interpolation, shown in Fig. 34
simplicity.
was chosen because of its
It is performed by linearly interpolating points along
separable orthogonal coordinates of the original 20 x 20 image field.
The displayed images are mapped into a pseudo color scale of
fifteen distinct colors from black to white.
For convenience,
wedge is displayed on the right hand side of each image.
a color
The color on
the top represents the largest value or the highest attenuation and
then it comes down the scale. The bottom one represents the smallest
value or
the
lowest attenuation.
implemented in the system.
Two color coding schemes were
They are stored in a data file "13.DATA"
and are read into the program when the display routine is activated.
Modification can easily done by modifying the "13.DATA" file prior to
the execution of the image processing program ACQ4.
92
\
Fig. 34 Two dimensional bilinear interpolation (borrowed from
Pratt, 1979)
CHAPTER 5
METHOD OF EXPERIMENT
For the purpose of testing the performance of the system described
in chapters 3 and 4 of this thesis, four experiments were performed.
Each study was undertaken with a specific objective in mind:
Experiment one -
to compare the actual sizes of test tubes with
that from the images.
Experiment two -
to compare different biological materials.
Experiment three -
to
find
objects
the
that
minimum
can
spacing
still
between
be
two
detected
distinctively.
Experiment four -
to observe the diffraction pattern.
This chapter is divided into two sections: (1) phantoms, and (2)
experimental protocol.
5.1 Phantoms
Thin-walled glass test tubes of varying sizes filled with various
biological materials were used as phantoms of these experiments.
the tubes are of thickness 0.1 cm and length
length of the transducer array.
acoustic attenuation of 13/m.
diameter sizes.
All
30 cm, covering the whole
They are made of glass and have
They are different only
in their
The diameter of each of the transducer elements
measures 0.9 cm, therefore, 0.9 cm was used as a unit to describe the
resolution of the system.
Three test tubes of diameter sizes, namely
0.6 cm (less than a unit), 0.9 cm (one unit), and 1.8 cm (2 units) were
chosen.
93
94
These tubes were filled with either 0.9 % saline, muscle phantom
(which was developed by the University of Washington; composition by
weight: 75.44% water, 8.45% supper stuff, 0.9069% salt and 15.2%
polyethylene powder), or glycol to simulate biological tissues with
high, moderate and low water content respectively.
These test tubes were held in stable positions by the use of semi­
circular
grooves
transducer array.
machined
on
the
same
plastic
rack
holding
the
With this holder, separations between test tubes
were maintained in fix increments of either one unit, two units, or
three units and the relative position between the transducer array and
the phantoms was known for all experiments.
5.2 Experimental Protocol
The experimental configuration was best described in Fig. 35.
In
all the experiments, microwave energy was delivered by an Elmed-107
applicator into the surface of a 25 x 30 x 50 cm tank of water. The
Elmed-107
applicator
has
a
diameter
of
30
cm.
It
thermoelastic waves at the contact of the water surface.
diverged
outward
as
spherical
waves.
The
matching
generated
The waves
between
the
microwave pulse generator and the applicator was carefully tuned by a
double stub tuner until the reflected power reduced to minimum.
pulse width was selected from the Tektronix pulse generator.
voltage was adjusted to 3 KV.
A 2 /is
The peak
The forward and reflected power were
monitored so that the peak power was about 30 KW. The hydrophone array
was placed on a plastic rack specially designed for adjusting the
horizontal level of the transducer and the relative position between
the applicator and the transducer.
It was designed to position the
TRANSDUCER OUTPUT
SIGNALS TO DAS
MICROWAVE
APPLICATOR
PULSED MICROWAVE
ENERGY
TANK FILLED
WITH WATER
OBJECT TO BE IMAGED
20 X 20
TRANSDUCER ARRAY
\
Fig. 35 Simplified diagram of the microwave-induced acoustical
imaging apparatus (borrowed from Olsen, 1982)
96
transducer array so that it was in the far field region from the
microwave source and the microwave energy reaching the phantom was
negligible.
According to Michaelson and Lin (1987), the penetration
depth for 2450 MHz microwave in water is about 1 cm.
Therefore, the
array was centered below the applicator at a depth of 14 cm in the
water,
satisfying
consisted
all
the
of irradiating
above
conditions.
These
experiments
the water with individual RF pulses and
measuring the initial peak response of each array element by the
system.
The amplified, filtered, and digitized signals were stored in
the computer and referred as "BACKGROUND IMAGE" throughout the thesis.
Another series of irradiations with the phantoms in place produced a
second set of data referred as "OBJECT IMAGE".
"BACKGROUND IMAGE" was subtracted from it corresponding "OBJECT
IMAGE".
The difference image was then normalized by the background
image pixel by pixel to eliminate the influence of non-uniform RF
illumination of the array and non-uniform gain throughout all channels.
The difference image was convolved with a 3 x 3 weighted -average low
pass
filter (see
Fig.
29) to
suppress high-frequency
noise.
A
threshold applicator was then applied to establish the boundary of the
test tubes.
Finally, the image was displayed on a color monitor with
15 distinct colors.
enhance
this
Histogram equalization can also be applied to
threshold
image.
Image
and
its histogram
at
each
intermediate step were displayed.
For the first experiment, a single test tube of 0.9 cm filled with
muscle-equivalent material was placed on top of the 10th row of the
transducer array.
The number of pixels occupied by the cross section
97
of the object on the thresholded image was taken as the size of the
test tube .
tube.
This number was compared with the actual size of the test
The same procedure was repeated with a 0.6 cm test tube at the
10th row of the array and a 1.8 cm test tube at the 10th and 11th rows
of the array.
The second experiment consisted of two test tubes of 0.9 cm, one
filled with muscle-equivalent material and the other with glycol set on
a plastic holder with a separation of 2.7 cm (3 units) (Fig. 36a).
The
positions of the two test tubes were symmetric with respect to the
center of the transducer array , as well as the applicator.
configuration,
these
two
tubes
received
equal
In such a
amount of acoustic
illumination.
The separation of 2.7 cm was designed to ensure minimum
interference.
The two objects from the threshold image were compared
in terms of their intensity distribution.
The procedure was repeated
with two same size tubes, one filled with muscle-equivalent material
and the other with 0.9% saline.
For the purpose of testing the resolution of the system, three
test tubes of 0.9 cm all filled with muscle-equivalent material were
arranged on a plastic holder with separations of 0.9 cm and 1.8 cm
(Fig.
36b).
The
resolution
was
then
measured
as
the
smallest
separation at which two objects could be placed and still be detected
distinctively.
Finally, diffraction effect on microwave-induced acoustic waves
was studied.
The transmission projection was obtained from the system
assuming that the ultrasound energy travelled in a straight line.
This
assumption simplifies the correction algorithm, which will be discussed
3.6
CM
Fig. 36 Experimental protocols: (a) two test tubes;
(b) three test tubes
99
chapter 6.
However, errors existed and diffraction of the ultrasound
energy did occur.
Greenleaf (1982) suggested a simple experiment to
observe this effect and correct the error.
A similar experiment was
devised to observe this effect for this system, but little effort has
been made to correct this error.
A single test tube of 0.9 cm filled
with air was placed on top of the 10th row of the transducer array.
To
prevent any movement, the test tube was secured to the hydrophone array
with a piece of string.
with
the
The processing was stopped after convolving
digital filter.
Air
is
a poor
conducting
medium
for
ultrasound energy. No transmitted acoustic waves can pass through the
tube.
Therefore, the receiving pattern is due to the diffraction
effect alone.
CHAPTER 6
RESULTS AND DISCUSSION
The images obtained from the four experiments described in the
previous chapter are presented and discussed in this chapter.
6.1 The Microwave-induced Acoustic Source
As described in chapter 4, the acoustic wave front generated at
the surface of the water tank does not distribute evenly throughout the
entire transducer array, but rather diverges outward from the center as
a spherical wave.
identical.
The gains on all signal processing channels are not
The typical amplitude background image of Fig. 37 shows
very clearly these effects.
In order to further describe the source function, a time-of flight
image was recorded.
Signals coming out from the S/H channels were fed
into a digital scope Nicolet 4094A.
The signals were sampled at 2 MHz
and the time from the starting point to the beginning of the first
response of each channel was measured.
These 400 data were then keyed
into the Intel 86/380 for further processing.
Figure 38 is a time-of-
flight image showing the time for the acoustic energy to propagate to
the entire transducer array.
This is only a crude image to illustrate
the source function in the time domain.
The adjacent transducer
element is 0.9 cm and the distance between the source to the transducer
array is 14 cm.
Assuming that the acoustic velocity in water is 1540
m/s, time difference from element to element is 0.2 /is.
rate is inadequate for the application.
100
The sampling
Also, the display system has
101
,
saaH£8^fe '"••••
8 •
s fiji
•; . •. m •;
V -s» '•
•.•••<, W
a \,,
• s»v.-a ,•
Fig. 37 An amplitude "BACKGROUND" image
• •
102
4
Fig. 38 A time-of flight "BACKGROUND" image
103
only 15 distinct colors to describe the differences.
It is also not
sufficient for the application.
The gain effect is very well demonstrated in Fig. 39.
These two
waves were coming from the same transducer, but going through different
processing channels.
The amplitude of the first wave is about 3.25 V
peak-to-peak, but that of the second wave is about 2.48 V peak-to-peak.
Of course, they may represent the extreme cases.
Nevertheless, this
problem does exist and the normalization seems to reduce this effect to
a great extent.
6.2 A Model
When the object was present, the wave was disturbed and deflected
away from the object (Fig. 40).
From this image, we attempted to
propose a model explaining these experimental observations.
illustrates this model.
Figure 41
We assume that the wave propagated away from
the applicator and travelled towards the transducer array in a straight
path.
When the object was present, it compressed the waves away from
the object creating a higher pressure at their interfaces (Fig. 41b).
This phenomenon may be related to the diffraction effect.
that travelled
The wave
through the tube was attenuated by the biological
material inside, which had a higher acoustic attenuation coefficient
than that of water (Fig. 41c).
The material inside was homogeneous.
The difference in intensity represented the different thickness of the
tubes.
By superposition, we obtain the profile in Fig. 41d.
When we
subtract the "OBJECT" image from the "BACKGROUND" image, we obtain a
profile of the form shown in Fig. 41e.
104
Fig. 39 Non-uniform gain of processing channels
105
Fig. 40 An "OBJECT" image: a 1.8 cm glass tube filled with muscle
equivalent material
106
(a)
(b)
(c)
(d)
(e)
Fig. 41 A model, (a) a cross-sectional view of a test tube filled with
homogeneous material;
(b) deflected wave at the interface;
(c) attenuation profile;
(d) (b) + (c);
(e) "BACKGROUND' image - "OBJECT" image.
107
6.3 Sinple Test Tube
Figure 42, 43, and 44 are processed images for single test tubes
of 0.6 cm, 0.9 cm and 1.8 cm, respectively. In Fig. 43 and 44 the test
tubes are clearly seen extending from one end of the transducer array
to the other end.
From a cross-sectional view, it is noticed that the
intensities are the highest in the middle of the tubes and attenuate
symmetrically toward both edges.
The intensities go further down and
come back up toward the top and bottom of the picture.
They are well
correlated with the physical structures of the tubes and the proposed
model described in the last section.
are assumed to be homogeneous.
the
different
attenuation
The materials inside the tubes
The difference in intensity represents
throughout
the
tubes,
which
corresponds to different thickness of the tubes (Fig. 45).
values can easily be applied to threshold these images.
in
turn,
Threshold
In Fig. 46 and
Fig. 47, all the below thresholded grey levels were compressed to the
first level and those above the threshold were stretched throughout the
rest of grey levels available.
Details can be seen more clearly in
these images than those before thresholding.
From the cross-sectional view of Fig. 42, the intensity also
correlates well with the physical structure of the tube and the model
explained before.
However, only the middle section of the test tube in
Fig. 42 is visible in the image.
It is very difficult to assign a
global threshold to separate the image into "object" and "background".
Several problems may account for this to take place:
(1) Signal-to-noise is lower than the other two images;
108
42 A filtered image: a 0.6 cm glass tube filled with muscle
equivalent material
109
43 A filtered image: a 0.9 cm glass tube filled with muscle
equivalent material
110
44 A filtered image: a 1.8 cm glass tube filled with muscle
equivalent material
Ill
CEDSS-5ECTI0N
CE055-5eCT(0M OF
A TUBE WITH
H0M0GEMEOU5 MATEKIAL5
PITTANCE
Fig. 45 The projection of a test tube
Fig. 46 A thresholded image: a 0.9 cm glass tube filled with
muscle-equivalent material
113
_^^^2LJ223^L
Fig. 47 A thresholded image: a 1.8 cm glass tube filled with
muscle-equivalent material
114
(2) The source concentrated at the center and fanned out evenly
to all directions.
The energy travelled to the side takes a
longer path than that to the center of the tube.
Therefore,
energy at the side is weaker due to acoustic attenuation.
(3) A 0.6 cm tube occupied 2/3 of a transducer.
It may very well
indicate a partial volume problem.
One way to compensate for these shortcomings is to apply a stronger
source to improve the signal-to-noise ratio and to use finer transducer
elements to reduce the partial volume problem.
In order
quantized
at
to
the
quantitatively
same
levels
evaluate
were
these
images, histograms
constructed (Fig.
48).
The
intensities of the 0.9 cm and 1.8 cm tube have most of their pixel grey
levels higher than 40.
The 0.6 cm tube has its pixel spreading all the
way to the first level.
The maximum intensity of this image is also
lower than the other two.
The number of pixels occupied by the cross-sections of the objects
in these thresholded images was computed and compared with the actual
sizes of the test tubes.
The results are presented in Fig. 49.
The
errors may be caused by the shadow of the tubes (see Fig. 50); by
blurring
introduced
diffraction
and
by
the
low-pass
filter
refraction
effects.
The
particularly observed in Fig. 52a and 52b.
(Fig.
51)
shadowing
;
or
by
effect
is
Figure 52a was taken with a
0.9 cm test tube 5 cm above the transducer while Fig. 52b was taken
with the same test tube on top of the transducer.
The tube in Fig. 52a
appears to be bigger in size because of this shadowing effect.
The
blurring especially at the edge by a low-pass filter is a well-known
40 -
Histogram Of Single Tube
Muscle Phantom In Water
(D = 0.6 cm, L = 30.0 cm)
30 -
20 -
10- '
0 - n_..
r?
pmn
wr? nrjM
%
n il
Histogram Of Single Tube
Muscle Phantom In Water
60 -
n
z>/
//
/
(D = 1.0 cm, L = 30-0 cm
/
/
/
40 -
7i
1
\
20 -
i
li
0-
50 -
Histogram Of Single Tuble
Muscle Phantom In Water
40
(D = 1.8 cm. L = 30.0 cm)
30 20
10 -
1
10
20
30
40
iiiiiiiii
50
60
Grey Level
Fig. 48 Histograms of muscle phantoms of sizes:
(top) D = 0.6 cm; (middle) D = 0.9 cm; (bottom) D = 1.8 cm
116
Diameter Of Test Tubes
3
Measured Size,pixels
2
5
T
0.6
0.8
1.4
1
Actual Size, cm
Fig. 49 Comparing the actual size of the test tubes to that
measured form the images
AlSWATE&
—>
117
r
50U&CE.
PHAMT0^
TRAMDUC£(^ PLAME.
Fig. 50 Errors from shadowing
(a)
(b)
Fig. 51 (a) An ideal edge; (b) after moving average filtering
(a)
(b)
52 An filtered image of a 0.9 cm test tube filled with muscleequivalent material, (a) placed 5 cm above the transducer;
(b) placed on top of the transducer array
120
problem.
A median filter
blurring effect.
median filters.
is under
investigation to reduce
this
Huang (1981) explains in detail the implementation of
The diffraction effect will be discussed later.
6.4 Biological Materials
Figure 53 is an image of two 0.9 cm test tubes, one filled with
muscle-equivalent material and the other with glycol.
distributions from both materials are very similar.
image of the same test tubes
The intensity
Figure 54 is an
, one filled with muscle-equivalent
material and the other with 0.9% saline.
The intensity distribution
from the muscle is much stronger than that from 0.9% saline.
These are
well predicted by Table III and the following equation:
Ia
= IaOe_Q!aX •>
(6.1)
where Ia is the acoustic energy at x m and IaQ is the energy at the
r
surface of the tube (x=0).
The attenuation at 30 KHz across a 0.9 cm
test tubes of 0.9% saline (= water), muscle-equivalent material (= soft
tissue) and glycol (~ fat) is estimated to be 2.025E-7, 6.75E-3, and
3.51E-3 respectively.
According to equation (6.1), taking IQ as the
acoustic energy just before arriving at the test tubes and I as the
energy just after passing through the tube, IWater> ^muscle>
are IQ, 0.993 IQ, 0.996 IQ.
an<l
*fat
Therefore, the intensity distribution from
muscle is similar to that from glycol, but stronger than that from 0.9%
saline.
When
the
acoustic
frequency
distribution changes significantly.
increases,
has
materials.
the
capability
of
intensity
Images of materials with different
water contents may be better differentiated.
thus
the
Nevertheless, the system
discriminating
different
Further researches need to be investigated.
biological
Fig. 53 A filtered image of two 0.9 cm glass tubes filled with
(top)
muscle-equivalent material;
(bottom) glycol
122
Fig. 54 A filtered image of two 0.9 cm glass tubes filled with
(top)
muscle-equivalent material;
(bottom) 0.9 % saline
123
6.5 Spatial Resolution
Experiment three was performed to test the spatial resolution of
the system.
Figure 55 is an image obtained by placing three 0.9 cm
test tubes of muscle-equivalent material 0.9 cm and 1.8 cm apart.
is clearly seen that the separation of 0.9 cm is not detectable.
finding is consistent with that from the first experiment.
It
This
In addition
to those problems explained in section 6.3, namely shadowing, blurring,
diffraction and refraction problems, waves scattered by two test tubes
also
interfere
with
each
other (Greenleaf
et
al.
1982).
This
interference may be constructive when two waves are in phase and
destructive when they are out of phase.
It is a very complicated
mathematical problem, and is beyond the scope of this thesis to explain
this phenomenon.
From this experiment one can conclude that the
smallest separation at which two 0.9 cm test tubes must be placed and
still be detected is more than one transducer unit.
6.6 Diffraction Pattern
Figure 56 is an image of a tube filled with air.
Since air is a
bad conducting medium for acoustic energy, this may represent the
diffraction pattern of the tube, which is very much related to the
shape of the object to be imaged.
Figure 57 shows the histogram of
this image compared with that filled with muscle.
wider spread than that for muscle.
This histogram has a
This experimental results agrees
with the traditional transmission ultrasonic imaging (Holbrooke et al.,
1974) in that the effects of both absolute and differential acoustic
absorption are minimal compared with the edge effects of refraction,
diffraction and reflection especially in this low frequency range.
124
*
I
•
Hi''
>•
<Y*5
v
_
u
>.«
f
*' •_
Pa
*K\
sKsa&iTctt*
_ {^
-. •
55 A filtered image of three 0.9 cm glass tubes filled with
muscle-equivalent material and placed 0.9 cm and 1.8 cm
apart
t
126
80
Histogram Of Single Tube
Muscle Phantom In Water
60
(D = 1.0 cm, L = 30.0 cm
r£]
J;
40
20
0
Histogram Of Single Tube
Air Phantom In Water
40
(D - 1.0 cm, L = 30.0 em)
30
20
10
r.Iliiill
"0 R
0
1
10
20
30
n
40
50
60
Grey Level
Fig. 57 Histograms of images:
(top)
muscle phantom;
(bottom) air
127
6.7 Time-of-Flipht Images
Time-of-flight images were obtained from feeding S/H outputs to a
digital storage scope, Nicolet 4094A one by one.
sampled at. 2 MHz.
The signal was
The time from the starting point to the first
acoustic wave arrival was measured for each channel.
These data were
then transferred to the Intel 86380 for processing.
Figure 58 is a
time-of-flight
image
equivalent material.
Fig. 39.
of
a 0.9 cm
test
tube
filled
with
muscle-
This image is noisier than its amplitude image in
This should not be considered as a reflection on the
potential use of these techniques, but rather it is an indication of
the
performance
of
our
system
and
equipment
at
this
time.
Specifically, sampling rate of 2 MHz is too low for this measurement.
Human errors were involved by taking those time measurement on the
scope.
In fact, from the traditional transmission ultrasound imaging,
time-of-flight images are found to be less sensitive to reflection and
refraction as attenuation imaging (Carson et al., 1977).
6.8 Attenuation Correction
As described in section 5.2, it was designed that each raw image
of these experiments would represent the initial peak responses of the
array elements.
A single transducer located at the side of the array
was used to signal the computer the arrival of these acoustic waves.
All
25
sample-and-holds would
digitization.
be
placed
into
the
hold
mode
for
The problem arises when these waves do not arrive at the
transducer array simultaneously.
A time-of-flight image of the source
(no phantom) alone was taken and
computer as a reference.
transferred
to
the Intel 86380
The time difference from each element to that
128
Fig. 58 A time-of-flight image: a 0.9 cm glass tube filled with
muscle-equivalent material
129
used for triggering was calculated.
Attenuation due to this time
difference was estimated and added to both "BACKGROUND" and "OBJECT"
images for correction.
Even with this attenuation correction, the problem may not be
always corrected.
Recording each channel at a sampling rate of 10 MHz
for 30 fj.s, 15 fis before and after the expected arrival time, is
proposed
as
an alternative way
to
solve
this
problem.
Pattern
recognition routine can easily be written to track the first peak of
this wave and measure both the time and amplitude at that location.
As
expected, very fast A/D and controlling microcomputer for digital
conversion and hugh memories (7.5 Kbytes of 25 channels) will be
required.
With the present technology, this will not be a difficult
and expensive task to complete.
6.9 False Triggering
The initial artifact of each signal (Fig. 18) and noise which came
through the processing channel may occasionally cause a false wave
arrival information to the computer.
In order to avoid the initial
artifact and false triggering by noise, the circuit would ignore the
first 80 fis after the process began and the circuit would not retrigger
until all 25 channels were digitized and stored in the memories.
This
80 fis delay time is adjustable through a potentiometer on the front
panel.
It is desirable to adjust this potentiometer from time to time
and observe the signal on a scope to make sure that the sample-andholds have caught the right signals.
130
6.10 Quality Control
Electronic elements on the interfacing circuits may occasionally
fail without notice.
A "CHECK" routine was designed and implemented in
the system to ensure that good quality images could be received.
It
simply force values in a data file to fall within a reasonable range.
A
will be displayed on the terminal to indicate a valid value and a
'?' to indicate a questionable one.
The program also replaces the
questionable ones by an average value of its 4 immediate neighbors.
This correction may not solve the problem, but it will serve as a
warning to the operator.
The operator may decide to retake the image.
If consecutive unsatisfactory images are obtained, service for the
system is suggested.
CHAPTER 7
CONCLUSION
This thesis is concerned with the development of a microwave.induced thermoelastic tissue imaging system.
the
data
acquisition
and
image
This development allows
processing
of
two-dimensional
thermoelastic tissue images by a microcomputer, which was heretofore
done manually and serially on an oscilloscope.
Attempts have been made
to demonstrate the performance of the system and the potential use of
this imaging modality.
Results from these studies agree well with findings previously
reported
by
other
researchers
both theoretically and empirically.
Currently, the system can detect objects of about 1 cm in size with
minimum separation of about 2 cm from each other.
Biological materials
with high, moderate, and low water contents can be differentiated from
resulting images.
over-emphasize
the
However, these results can not be considered to
optimum
performance
of
this
imaging
modality.
Suffice it to say, these results have revealed that the designed
microwave-induced thermoelastic tissue imaging system is encouraging.
Nevertheless, the system has its own limitations and drawbacks.
Image processing capability of the system enhances the visibility
of useful information of these images, and in turn helps to identify
the
phantoms
from
these
processed
images.
The
flexibility
and
expendability of this software will accommodate further improvement on
the hardware of the system.
131
132
From literature in ultrasonic transmission tomography, time-offlight images are proven to be less distorted than the amplitude
images.
the
An attempt has been made to investigate this method.
limitation
obtained.
of
the
equipment
available,
noisier
Due to
images
were
Conclusion on this technique can not be further drawn.
While the system has been primarily tested on artificial simple
phantoms, animal and human tissues are categorized into high, moderate,
and low water contents materials.
It is therefore believed that this
imaging modality can be useful clinically for tissue classification in
the future.
The main objective of this thesis is to build a prototype system
to investigate the feasibility of using microwave-induced thermoelastic
waves as an imaging modality.
means complete.
The design of this prototype is by no
Many areas are still in need of further investigation.
These include:
(1) Investigation
of
an
optimum
combination
of
microwave
frequency and pulse widths for tissue imaging.
(2) Design of a narrow band-limited high frequency transducer
array with possible higher spatial resolution.
(3) Design a
data
acquisition system
that can acquire both
amplitude and time-of-flight images.
Pattern recognition
routines are in need to accurately track the beginning of the
wave arrival.
(4) Investigation
of
the
possible
advantages
attained
by
implementation of other filter realizations (e.g. median
filter).
133
(5) Testing the system with multi-layer phantoms with various
shapes.
(6) Corrections on diffraction and refraction effects.
In conclusion, this thesis has demonstrated the feasibility of
using microwave-induced thermoelastic waves as an imaging modality.
This prototype offers an excellent tool to researchers for further
investigation.
BIBLIOGRAPHY
Analog Devices: Data-Acquisition Data book, vol 1, Norwood, Analog
Devices, 1982.
Applied Microwave: Microwave Pulsed Signal Source-Model PH40k Manual,
Westwood, Applied Microwave, Inc., 1976.
Bates, R.H.T., Garden, K.L., and Peters, T.M.: Overview of computerized
tomography with emphasis on future developments. Proc. IEEE,
71:356-372, 1983.
Borth, D.E., and Cain, C.A.: Theoretical analysis of acoustic signal
generation in materials irradiated with microwave energy. IEEE
Trans. Microwave Theory Tech., MTT-25:944-954,1977.
Bowen, T.: Radiation-induced thermoacoustic soft tissue imaging,in
Ultrasonics Symposium, 1:817-822, 1981.
Bowen, T.,Nasoni, R.L., and Pifer, A.E., and Sembroski: Some
experimental results on the thermoaco.ustic imaging of tissue
equivalent phantom materials, in Ultrasonics Symposium, 1:817822,1981.
Bowen, T. , Nasoni, R.L., . and Pifer, A.E.: Thermoacoustic imaging
induced by deeply penetrating radiation. In Acoustic Imaging,eds.
Kaveh, M, Mueller,R.K., and Greenleaf, J.F., New York, Plenum
Press, 13:409-427, 1983a.
Bowen, T. , Connor, W.G., Nasoni, R.L., and Pifer, A.E., Bell, R. ,
Cooper, D.h., and Sembroski, G.H:: Thermoacoustic imaging induced
by deeply penetrating radiation. In Acoustic Imaging, eds. Kaveh,
M, Mueller,R.K., and Greenleaf, J.F., New York, Plenum Press,
13:429-434, 1983b.
Brownell, G.L., Budinger, P.C., Lanterbur, P.C., and McGreer, P.L.:
Positron tomography and nuclear magnetic resonance. Science,
215:619-626,1982.
Byrom,E., Pavel, D. , Swiryn, D. , and Meyer-Pavel, C.:Phase images of
gated cardiac studies: a standard evaluation procedure.
In
(Proc.) Functional Mapping Of Organ Systems and Other Computer
Topies, New York, Society of Nuclear Medicine, 130-136, 1981.
Carson, P. Oughton, T. Hendee, W. and Ahuja, A: Imaging soft tissue
through bone with ultrasound transmission tomography by
reconstruction, Med. Phy., 4: 302-309, 1977.
134
135
Caspers, F. and Conway. J.: measurement of power density in a lossy
material by means of electromagnetically induced acoustic signals
for non-invasive determination of spatial thermal absorption in
connection with pulsed hyperthermia.
Proc. 12th Eur. Microwave
conf. (Helsinki), 565-568,1982.
Cavailloles,F., Bazin, J.P., DiPaola, R.: Factor analysis in gated
cardiac studies, J. of Nucl. Med., 25:1067-1079, 1984.
Chin, R.T. and Yeh, C.L.: Quantitative evaluation of some edgepreserving noise-smoothing techniques, Computer Vision, Graphic
and Image Processing, 23:67-91, 1983.
Christensen, E.E., Curry, T.S. Ill and Nunnally, J.: An Introduction to
the Physics of Diagnostic Radiology, Philadelphia, Lea and
Febiger, 1972.
Cobbold, R.S.: Transducers for Biomedical Measurements: Principles and
Applications, New York, Wiley-Interscience 170-174, 1974.
Datel Intersil: Data Conversion Components, Mansfield, Datel Intersil,
Inc., 202-209,1983/84
Das, P.: Digital enhancement of ultrasonic bone images, Ultrasonics
Symposium Proc. ,2:718-722,1982.
Davis, L.S.: A survey of edge detection techniques, Computer
and Image Processing, 4:248-270, 1975.
graphics
Duda, R. and Hart, P: Pattern Classification And Scene Analysis, New
York, John Wiley & Sons, 1973
Frei, W.: Image enhancement by histogram hyperbolization, Computer,
Graphics and Image Processing, 6:286-294, 1977.
Foster, K.R. and Finch, E.D.: Microwave hearing:evidence for
thermoacoustical auditory stimulation by pulsed microwave.
Science, 185:256-258,1974.
Gournay, L.S.: Conversion of electromagnetic to acoustic energy by
surface heating. J. Acoust. Soc. Am., 40:1322-1330,1966.
Greenleaf, J.F. and Bahn, R.C.: Clinical imaging with transmissive
ultrasonic computerized tomography, IEEE Tran. on Biomed. Eng.,
28: 177-185, 1981.
Greenleaf, J.F., Gisvold, J.J., Bahn, R.C.: Computed Transmission
Ultrasound tomography, Med. Progr. Technol., 9:165-170, 1982a.
Greenleaf, J.F., Gisvold, J.J., Bahn, R.C.: A clinical prototype
ultrasonic transmission tomographic scanner, In Acoustical
Imaging, ed. , New York, Plenum, 12:579-587,1982b.
136
Greenleaf, J.F.: Effects of diffraction and refraction on computerassisted tomography with ultrasound, Proc. IEEE, 71:233-337, 1982
Guy, A.W., Chou, C.K., Lin, J.C. and Christensen, D.: Microwave-induced
acoustic effects in mammalian auditory systems and physical
materials, Ann N Y Acad Sci, 247:194-218, 1975.
Hall, E.L.: Computer Image Processing and Recognition, New York,
Academic Press, 1979.
Hendee, W.R.: Medical Radiation Physics, Chicago, Year Book Medical
Publishers, Inc., 1979.
Hentz, V., Marich, K., and Dev, P: Preliminary study of the upper limb
with the use of ultrasound transmission imaging, J. Hand Surgery,
9A: 188-193, 1984
Herman, G.T.: Image Reconstruction From Projections: the fundamentals
of computerized tomography, New York, Academic Press, 1980.
Hiller, D., Ermert, H.: Ultrasound computerized tomography using
transmission and reflection mode: application to medical
diagnosis. In Acoustic Imaging, ed. , New York, Plenum, 12:553587,1982.
Holbrooke, D. McCurry, E., Richards, V. and Shibata, H.: Trough
transmission ultrasonic imaging of intrauterine and fetal
structures using acoustical holography, Proc 2nd World Cong.
Ultrasonics in med., 332-344, 1974
Huang, T.S. (ed): Two-Dimensional Digital Signal Processing IITransforms And Median Filters, New York, Springer-Verlag, 1981.
Hummel, R. Histogram modification techniques, Computer, Graphics and
Image Processing, 4:209-224, 1975.
Hutchins, D.A. and Tam, A.C.: Pulsed photoacoustic materials
characterization, IEEE Trans. Ferroelec. Freq. Contr., 33:429-449,
1986a.
Hutchins, D.A. and Tam, A.C.: Special issue on photoacoustics,, IEEE
Trans. Ferroelec. Freq. Contr., 33:427-428, 1986b.
James, Jr., Price, R.R., Stewart, R., Partain, C.L., Harms, S.: Nuclear
magnetic resonance imaging: An overview, in Reba, et al. (eds),
Diagnostic Imaging in Medicine, Boston, Martinus Nijhoff
Pulblishers, 214-230, 1983.
137
Johnson, S.A., Greenleaf, J.F., Rajagopalan, B. and Bahn, R.C.:
Ultrasound images corrected for refraction and attenuation: A
comparison of new high resolution images, in Raviv, et al. (eds),
Computer Aided Tomography and Ultrasonics in Medicine, New York,
North-Holland Publishing Co., 1979.
Kak, A.C.: Special issue on computerized medical imaging, IEEE Trans.
Biomed. Eng., 28:49-234, 1981.
Kouris, K. , Spyrou, N.M. and Jackson, D.F. ed.: Imaging with Ionizing
Radiations, Lodon, Surrey Univ., 1982.
Leeman, S.: Ultrasound transmission and scattering in human tissue, in
McAinsh, T.F. (eds), Physics in Medicine and Biology Encyclopedia,
New York, Pergamon Press, 852-856, 1986.
Lin, J.C.: Microwave auditory effect-A comparison of some possible
transduction mechanisms, J Microwave Power, 1:77-81, 1976.
Lin, J.C.: On microwave-induced hearing sensation, IEEE Trans, on
Microwave Theory and Tech., 25:605-613, 1977a.
Lin, J.C.: Further studies on the microwave auditory effect, IEEE
Trans, on Microwave Theory and Tech., 25:938-943, 1977b.
Lin, J.C.: Microwave Auditory Effects and Applications, Springfield,
Charles C. Thomas, 1978.
Lin, J.C., Meltzer, R.J. and Redding, F.K.: Comparison of measured and
predicted characteristics of microwave-induced sound, Radio
Science, 17:159S-163S, 1982.
Lin, J.C. and Chan, K.H.: Microwave thermoelastic tissue imagingsystem design, IEEE Trans. Microwave Theory Tech., MTT-32:854860,1984.
Macovski, A.: Medical Imaging System, Englewood Cliffs, Prentice-Hall,
1987.
Marich, K. , Zatz, L. , Green, P., Suarez, J., and Macovski, A.: Real­
time imaging with a new ultrasonic camera: Part I, in vitro
experimental studies on transmission imaging of biological
structures, J. clin. Ultrasound, 3: 5-16, 1975.
Mastin, G.A.: Adaptive filters for digital image noise smoothing,
Computer Vision Graphics and Image Processing, 31:103-121, 1985.
Meltzer, R. , Sartorius, 0., Lancee, C. , Serruys, P., Verdouw, P.,
Essed, C., and Roelandt, J.: Transmission of ultrasonic contrast
through the lungs, Ultrasound in Med. and Biol, 7:377-384, 1981.
138
Michaelson, S.M. and Lin, J.C.: Biological Effects and Health
Implications of Radiofrequency Radiation, New York, Plenum Press,
1987.
Nasoni, R., Evanoff, G.A.,Jr. , Halverson, P.G. ,
Bowen, T.:
Thermoacoustic emission by deeply penetrating microwave radiation.
IEEE 1984 Ultrasonic Symposium Proceedings, 2:633-638,1983.
National Semiconductor: Linear
Semiconductor, Inc.1982.
Data
Book,
Sunnyvale,
National
Newton, T.H. and Potts, G. ed: Radiology of the Skull and Brains, St.
Louis, C.V. Mos, 1981.
Olsen, R.G.and Hammer, W.C.: Microwave-induced pressure waves in a
model of muscle tissue, Bioelectroinagn., 1:45-54, 1980
Olsen, R.G. and Lin, J.C.: Microwave pulse-induced acoustic resonances
in spherical head models, IEEE Trans, on Microwave Theory and
Tech., 29:1114-1117, 1981.
Olsen, R.G.: Generation of acoustic images from the absorption of
pulsed microwave energy. In Acoustic Imaging, ed. Powers, J.P.,
New York, Plenum, 11:53-59,1982.
Olsen, R.G. and Lin, J.C.: Acoustic imaging of a yiiodel of a human hand
using pulsed microwave irradiation, Bioelectromagn., 4:397400,1983.
Panda, D.: Nonlinear smoothing of pictures, Computer Graphic and Image
Processing, 8:259-270, 1978.
Panda, D.: Statistical properties of threshold images,
Graphics and Image Processing, 18:334-354, 1978.
Computer
Pavel, D., Byrom, E., Lam, W. , Meyer-Pavel, C. Swiryn, S,,Pietras, R.:
Detection and quantification of regional wall motion abnormalities
using phase analysis of equilibrium gated cardiac studies, Clin.
Nucl. Med., 8:315-321, 1983.
Peli, T. and Malah, D.: A study of edge detection algorithms, Computer
Graphic and Image Processing, 20:1-21, 1982.
Pratt, W.K.: Digital Image Processing, New York, John Wiley and Sons,
1978.
Rassow, : Basic information on routine diagnosis in nuclear medicine.
In Imaging For Medicine, ed. by Nudelman, S. and Patton, D.D., New
York, Plenum Press, 1:1980.
Rosenfeld, A. and Kak, A.C.:
Academic Press, 1,2:1982.
Digital Picture Processing, New York,
139
Schuy, S.: Medical Ultrasonic imaging, Med. Prog. Technol., 9:161164,1982.
Sharp, J.C., Grove, H.M., and Gandhi, O.P.: Generation of acoustic
signals by pulsed microwave energy. IEEE Trans. Microwave Theory
Tech. MTT-22:583-584,1974.
Shaw, D.: Nuclear magnetic resonance imaging, in McAinsh, T.F. (eds),
Physics in Medicine and Biology Encyclopedia, New York, Pergamon
Press, 529-544, 1986.
Siliconix: Small Signal Design Catalog, Santa Clara, Siliconix, Inc.,
5-27, 1982.
Su, J.L. and Lin, J.C.: Thermoelastic signatures of tissue phantom
absorption and thermal expansion. IEEE Trans. Biomed. Eng., BME34:179-182,1987.
Su, J.L.: Acoustic imaging of induced thermal expansion of simulated
biological tissue.
Master Thesis, University of Illinois at
Chicago, 1985.
Texas Instruments: Linear Data Book, Dallas, Texas Instruments, Inc.,
1985.
Weigel, J.P., Cartee, R.E., Marich, K.W.: Preliminary study on the use
of ultrasonic transmission imaging to evaluate the hip joint in
the immature dog, Ultrasound Med. And Biol., 9:371-378,1983.
Weszka, J.S.: A survey of threshold selection techniques, Computer
Graphics and Image Processing, 7:259-265, 1978.
White, R.M.: Generation of elastic waves by transient surface heating,
J. Appl. Physics, 34:3559-3567, 1963.
Zatz, L.M., Marich, K.W. , Green, P.S., Lipton, M, Suarez, J., and
Macovski, A.: Real time imaging with a new ultrasonic camera: Part
II, preliminary studies in normal adults, J. Clin Ultrasound,
3:17-22, 1975.
Zatz, L. M.: Initial clinical evaluation of a new ultrasonic camera,
Radiology, 117:399-404, 1975.
APPENDIX A
FREQUENCY SPECTRUM OF MICROWAVE-INDUCED
ACOUSTIC WAVES
141
FREQUENCY SPECTRUM OF
A MICROWAVE-INDUCED ACOUSTIC SIGNAL
2450 MHz, 0.5 /xS
"id3
io4
io5
io6
FREQUENCY , Hz
59 Frequency spectrum of acoustic wave induced by thermoelastic
mechanism in water irradiated with 2450 MHz, 0.5 ps microwave
pulse.
142
FREQUENCY SPECTRUM OF
A MICROWAVE-INDUCED ACOUSTIC SIGNAL
2450 MHz, 1.0 jllS
z:
•
t>T'o.
Eh—• ^
W
w
E55
3
z:
O.
1CP
10
10
icf
FREQUENCY , Hz
60 Frequency spectrum of acoustic wave induced by thermoelastic
mechanism in water irradiated with 2450 MHz, 1.0 /zs microwave
pulse.
143
FREQUENCY SPECTRUM OF
A MICROWAVE-INDUCED ACOUSTIC SIGNAL
2450 MHz, 5.0 /xS
FREQUENCY , Hz
61 Frequency spectrum of acoustic wave induced by thermoelastic
mechanism in water irradiated with 2450 MHz, 5.0 jus microwave
pulse.
144
FREQUENCY SPECTRUM OF
A MICROWAVE-INDUCED ACOUSTIC SIGNAL
2450 MHz, 10.0
10
FREQUENCY
. 62 Frequency spectrum of acoustic wave induced by thermoelastic
mechanism in water irradiated with 2450 MHz, 10.0 /zs microwave
pulse.
145
FREQUENCY SPECTRUM OF
A MICROWAVE-INDUCED ACOUSTIC SIGNAL
2450 MHz, 20.0 /xS
10
l
FREQUENCY , £
. 63 Frequency spectrum of acoustic wave induced by thermoelastic
mechanism in water irradiated with 2450 MHz, 20.0 ns microwave
pulse.
146
FREQUENCY SPECTRUM OF
A MICROWAVE-INDUCED ACOUSTIC SIGNAL
433 MHz, 2.0 /xS
/
iii
$
it?
10'
lCf
lrf
FREQUENCY , Hz
64 Frequency spectrum of acoustic wave induced by thermoelastic
mechanism in water irradiated with 433 MHz, 2.0 /zs microwave
pulse.
147
FREQUENCY SPECTRUM OF
A MICROWAVE-INDUCED ACOUSTIC SIGNAL
5000 MHz, 2.0 S
m
b
z:
x
>H O .
h-H
W
w
H
.7_
o.
lCf
10'
ioP
10
FREQUENCY , Hz
Fig. 65 Frequency spectrum of acoustic wave induced by thermoelastic
mechanism in water irradiated with 5000 MHz, 2.0 /zs microwave
pulse.
148
FREQUENCY SPECTRUM OF
A MICROWAVE-INDUCED ACOUSTIC SIGNAL
8000 MHz, 2.0 £iS
'o
.2.
;s:
•7"
>-Tb.
(—4
Z
W
EZ
.z.
"7^
'O.
id 3
ltf
ltf
10
FREQUENCY , Hz
Fig. 66 Frequency spectrum of acoustic wave induced by thermoelastic
mechanism in water irradiated with 8000 MHz, 2.0 f i s microwave
pulse.
APPENDIX B
SCHEMATIC DIAGRAMS OF THE DATA ACQUISITION SYSTEM
9 -H5V
I5v
H5Y
I? LP
CHP
-v
x x ^ n r v i 1i i
rT
Fl2oM
o - 15V
? -M5V
Tap T
r T
Arr^V
,v
t^x
T
T
* -I5V
AHPLIFIEE,
5TAG£ 1
Fig. 67 A schematic diagram of a passive filter
HI—i—il
pTTTfT'I
(?HP
i5V
I
6 -tsv
TO
AMPLIFIER
3TA6££
H>tbic-
E>E>-
VM
*WM-ICI
vW VnCAie
9 IsjZjfl
5f«V
i,
3t
5,
.
Tc AuALflfe
MOX
12* L£V£L)
DI61TAU COWT^DL
AO 75 46
* ia-CHAMUEL
2* AWAL06
25 .
57 MUX
It
OUT
Sfl
V<£, GAJD
£,«
"'-10
ToLiBC
5i7
5,i
5.2
S.s
VS5
3*
S,B
5*
C*iO
Vdp
EM A* A, AzAa
'S|t7
<+c4G&fF JL
n
©
Vu,G*&
IaO f^")-
-T4.
-v&
Of7>Via.6M^
e>•Vfi
0000
Fig. 68 A schematic diagram of the data acquisition system I
152
£6
AO 7506
16 CHANNEL
P+-I5V
ANALOG
Q+\bo(-«-5v)
MUX
I I4-T> TKIK\
pcwee
SUPPLY EETUKU
s/*
ZM
A\ fit.**
16 17 16 IS
CU. OW
CTi -4*
rrt-/i6
-44
?i-4Z
AO 7574
LF356 k^V
5,
31-36
m -349u<r C5 et3
53
5*
u AD 7506
$616 CHAMM6L
5v aajalo<S
&£5'4
2> A&5-5
3
8^55-a^
,r3<3*C>
MUX
74139
Hi
EM2. Ao A i Aa A3
16 i7 16
o*im5
© T{-ZD
±- 5TAer
3WITCK
V/./SA QSuc Qo
C<«ey *74-lQ3>
ABC06UD WAD
T{-i4J*!-©
Fig. 69 A schematic diagram of the data acquisition system II
APPENDIX C
PROGRAM LISTINGS
154
iBHX 86 Paacal-88, V2.U
Source File: ACQ4.PASC
Object File: ACQ4.OBJ
Controls Specified: <none>.
STMT
1
2
3
4
5
6
7
1
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
LINE
1
2
3
4
5
6
NESTING
0 0
0 0
0 0
0 0
0 0
0 0
7/ fl
U fl
U
8
9
10
ii
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
0
0
0
o
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
SOURCE TEXT: ACQ4.PASC
MODULE SYSTEH;
PUBLIC CTLJSHJ;
PROCEDURE INJATAJ(VAR PICTURE:INTEGER);
PUBLIC CTLJSHJ;
PROCEDURE INJATAJ(VAR PICTURE:INTEGER;J:INTEGER);
PUBLIC IPJASCJ;
PDnnpnrTRF
?DAPI?.
rflUOBl/Uflli AUI
flVUKfluli,
PUBLIC IP PASC 2;
PROCEDURE DISPLAY 1
PROCEDORE DISPLAY 2
PROCEDURE DISPLAY 3
PROCEDURE DISPLAY 4
PROCEDURE DISPLAY 5
PROCEDURE DISPLAY 6
PUBLIC DISPLAY;
PROCEDURE INIT IHAGE;
PROCEDURE CLEARJSPY;
PROCEDURE DSPYJHAGEIVAR DISPLAY:IHTKGEB);
PROCEDURE DSPYJRRAY(VAR LEVEL,DISPLAY:INTEGER);
PROCEDURE DSPY CODE
PUBLIC IP PASC 3;
PROCEDURE ENHANCE;
PROCEDURE HISTR;
PROCEDURE IISTKVAR PICTURE:IHATRIX1);
PUBLIC IPJASC 4;
PROCEDURE SUBTRACT;
PUBLIC IP PASC 5;
PROCEDURE THRESHOLD;
PUBLIC IPJASC 6;
PROCEDURE GREYT;
PROCEDURE COHVERTI;
PROCEDURE C0NVERI2;
PROCEDURE CONVERTR;
PROCEDURE EQUAL;
PUBLIC IPJASC 7;
PROCEDURE CONVOLVE;
PUBLIC IPJASC 8;
PROCEDURE CHECK;
PUBLIC SYSTEH;
TYPE RHATRIX=ARRAY [1..20,1..20] OP REAL;
RHATRIX2=ARRAY[1..20]OF REAL;
IHATRIX1--ARRAY [1..20,!..20] OP INTEGER;
IHATRIX2-ARRAY [1..256] OF INTEGER;
44
45
46
47
48
49
50
51
44 0
45 0
46 0
47 0
48 0
49 0
50 0
51 0
0
0
0
0
0
0
0
0
IHATRIX3=ABRAY[l..ll)]0f INTEGER;
IHATRIX4=ARRAY [1.. 6 0 , 1 . .6010F INTEGER;
VAR IN_1:TEXT;
PIGJO:PAGKED ARRAY[1..13] OF CHAR;
NAME:ARRAY [1..13] OF CHAR;
T256:ARRAY [1..256] OF REAL;
ROW,COLOHN,LEVEL1,LEVEL2,LEVEL,CODE:INTEGER;
ANSHER:CHAR;
iRHX 86 Pascal-86, V2.0
STHT LINE RESTING
52 52 0 0
53 53 0 0
54 54 0 0
55 55 0 0
56 56 0 0
57 57 0 0
58 58 0 0
59 59 0 0
60 60 0 0
61 61 0 0
61
62
63
63 0 1
64 0 1
65 0 1
64
66
68
67 0 1
68 0 1
69 0 1
69
70
71
72
73
74
75
76
77
78
79
80
71
72
73
74
75
76
77
78
79
80
81
82
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
SOORCE TEXT: ACQ4.PASC
PITRE:RHATRIX;
IPITRE1,IPITRE2:IHATRIX1;
HISTO:IHATRIX2;
PROFILE:RHATRIX2;
DSPY:IHATRIX4;
PROGRAH SYSTEH( INPUT,OOTPtJT);
LABEL 5,10,14,15,20,30,40,50,60,70;
VAR PICTORE:ARRAYf 1..400] OF INTEGER;
PROCESS,I,J,K,L,H:INTEGER;
(* —
BEGINNING OF THE PROGRAH
*)
BEGIN
IHIT IHAGE;
CLEAR DSPY;
(* —
BUILDING THE TABLE
*)
T256[128]:=0.0;
FOR I: = 127 DOWNTO 1 DO T256[I]:-T256[I+l]-0.078;
FOR I:=129 TO 256 DO T256[I]:=T256[I-1]+0.078;
(* —
5:
—
*)
HRITELN('PLEASE ENTER THE DIMENSION OF THE PICTORE
HRITK('ROH=');
READLN(ROH);
HRITECCOLOHN:');
READLN(COLOHN);
HRITELNC 1. DATA ACQUISITION');
HRITELNC 2. IHAGE PROCESSING');
HRITELNC 3. DISPLAY');
HRITELNC 4. CHECK');
HRITELNC 5. EXIT');
HRITEC PLEASE ENTER A NUMBER —-> ');
READLN(PROCESS);
CASE PROCESS OF
156
81
83 0 2
82
0 2
83
84
85
85 0 2
87 0 2
86
88 0 2
88
90
91
89 0 2
90 0
91 0
92
93
94
95
96
97
99
93
94
95
96
97
98
99
0 2
0
0
0
0
0
0
0
(*
10:
I:-1;
14:
100 100 0
101 101 0
103
104
105
106
107
102
103
104
105
106
1: GOTO 10;
2: GOTO 40;
3: GOTO 50;
4: GOTO 60;
5: GOTO 70;
OTHERWISE HRITELN('ERROR — TRY AGAIN!');
END;
GOTO 5;
DATA ACQUISITION — *)
WRITELIH 'READY???');
HRITELtH'PRESS THE START BUTTON TO START-—)');
0
0
0
0
0
(»
J:= l;
IN JATA_2(PICTORE[ I]);
I:=1+25;
IF (I >=ROH*COLOHN)THEH GOTO 15;
HRITEC I=',I,' : CONTINUE?—>Y/N ');
READLN(ANSWER);
IF (ANSWER = 'N')THEN GOTO 15;
J:=J+1;
WRITELNC PRESS THE START BUTTON TO START—)');
IH JATA _3(PICTDRE[ I
GOTO 14;
REARRANGE THE DATA INTO A 20X20 ARRAY
iSHX 86 Pascal-86, 72.
STHT LINE NESTING
108
109
109
110
110
111
111
112
112
113
114
116
118
119
121
122
12!
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
2
2
3
3
4
4
5
5
5
4
3
3
2
2
1
SOURCE TEXT: ACQ4.PASC
15:
H:= 0;
J:=0;
REPEAT
I :=0;
REPEAT
FOR K:=J+1 TO J+5 DO
BEGIN
FOR L:=1+1 TO 1+5 DO
BEGIN
H:=H+1;
PITRE[L,K]:=T256[PICT0RE[H]+1];
END;
END;
I:=1+5;
UNTIL I>=ROW;
J:=J+5;
UNTIL J>=COLUUN;
*)
157
(*
125
126
128
129
130
132
133
134
135
137
138
139
140
140
141
142
144
145
146
147
148
150
151
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
1
1
1
1
1
1
1
152
153
154
155
156
157
158
159
160
161
162
163
150
151
152
153
154
155
156
157
158
159
160
161
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
2
2
2
2
20:
30:
(*
40:
SAVE THE PICTURE
*)
WRITEf'DO YOD HAHT TO SAVE THE PICTURE?—>Y/8 ');
READLNfANSWER);
IF (ANSWER =T)THEH GOTO 20;
HRITEf'DO YOO HAHT TO RETAKE THE PICTDBE?—>Y/B ');
READLN(ANSWER);
IF (ANSWER =T) THEN GOTO 10;
GOTO 30;
HRITELN('PLEASE ENTER THE FILE NAHE FOR STORING THE PICTORE —> :PBOG:PX.YYYY*
FOR I:=l TO 13 DO
IF NOT EOLN THEN READ(NAHE[I]);
PACK(NAHE, 1, PICJO);
REHRITE(IN_l,PICJO);
FOR I:=l TO ROW DO
BEGIN
FOR K:=l TO COLUMN DO
WRITE(INJ,PITRE[I,K]:12);
END;
WRITELN(INJ);
READLN;
WRITE('HORE IMAGES?
>Y/N ');
READLN(ANSHER);
IF (ANSWER = 'Y') THEN GOTO 10;
GOTO 5;
IMAGE PROCESSING
*)
HRITELNC(*
—- IMAGE PROCESSING
»)');
HRITELNC 1. AVERAGING ');
HRITELNC 2. ENHANCEMENT');
HRITELNC 3. SUBTRACTION WITH NORMALIZATION'):
HRITELNC 4. THRESHOLDING');
HRITELNC 5. EXIT');
HRITEf 'PLEASE ENTER A NOHBER -—>');
READLN(PROCESS);
CASE PROCESS OF
1: AVERAGE;
2: ENHANCE;
3: SUBTRACT;
4: THRESHOLD;
iRHX 86 Pascal-86, V2.0
STMT LINE NESTING
164 162 0 2
165 163 0 2
01/2
SOURCE TEXT: ACQ4.PASC
5: GOTO 5;
OTHERHISE HRITELNC ERROR — TRY AGAIN!');
167
169
170
171
172
173
164
165
166
167
168
169
1 7 4 no
175 171
0
0
0
0
0
0
0
0
2
1
1
1
1
1
1
1
50:
60:
70:
END;
GOTO 40;
DISPLAYJ;
GOTO 5;
CHECK;
GOTO 5;
HRITELH('********** BYE BYE! mm****')
END.
Summary Information:
PROCEDURE
SYSTEM
-COHST IN CODE-
OFFSET CODE SIZE
DATA SIZE
STACK SIZE
02D4H 06B0H 1712D 3277H 12919D OOOEH 14D
02D4H 724D
Total
171 Lines Bead.
0 Errors Detected.
495! Utilization of Hemory.
0984H 2436D 3277H 12919D 0042H
66D
8086/87/88/186 MACRO ASSEMBLER
CTLJSHJ
iEHX 86 8086/87/88/186 MACRO ASSEMBLER V2.0 ASSEMBLY OF MODULE CTL ASH_2
OBJECT MODOLE PLACED IH :FDD:CONTROL J.OBJ
ASSEMBLER INVOKED BY: :LANG:ASH86 :FDD:CONTROL_l.ASH
LOC OBJ
LINE
DATA ACQUISITION ROUTINE
AUTHOR: K.H. CHAN
DATE: 1/14/88
THIS ROOTINE CONTROLS ALL iHE I/O
FLIP-FLOPS
IT INTIALIZES THE
1
2
3
4
5
6
CTLJSHJ
7
8
9
10
0060
0050
0061
0010
0001
0002
0004
0019
00B5
00B6
0008
00B4
00B7
0099
0000
0000
0001
0002
0003
0004
0005
0006
5A
59
5B
58
51
52
8EC0
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
INIT1
INIT2
INIT3
PULS1
BOSY
DONE
START BIT
FINISH
PORTO
PORT1
ST CONVRT
A JO D
CTLJEG
CTL HOED
DATA ENDS
EQO
EQO
EQO
EQU
EQO
EQO
EQO
EQO
EQO
EQO
EQU
EQU
EQU
EQU
26
;
27
28
29
30
31
32
33
34
35
36
DATAJN SEGMENT POBLIC
POBLIC IN_DATA_2
INJATAJ PROC FAR
POP DX
POP CX
POP BX
POP AX
POSH CX
POSH DX
MOV ES.AX
60H
50H
61H
10H
01H
02H
04H
19H
0B5H
0B6H
08H
0B4H
0B7H
99H
IE
55
000A B91900
OOOD B099
000F E6B7
0011
0013
0015
0017
B060
E6B5
B050
E6B5
37
POSH DS
38
POSH BP
39
ASSUHE CS: DATA JH.DS:DATA
40
MOV CX,FINISH
41
;
42
;SET THE I/O POSTS DATA FLOW DIRECTION
43
HOV AL.CTLJORD
44
ODT CTLJEG,At
45
;INITIALIZING THE CODNTERS
46
HOV AL.IHITl
47
OUT PORTO,AL
48 '
HOV AL.INIT2
49
OCT PORTO,AL
50
;
8086/87/88/186 MACRO ASSEHBLER
LOC OBJ
0019 E4B6
001B 2404
001D 75FA
001F B010
0021 E6B5
0023 E4B6
0025 2408
0027 74FA
0029
002B
002D
002F
0031
0033
E4B6
2401
74FA
E4B4
E6B4
90
0034 E4B6
0036 2401
0038 74FA
003A E4B4
003C F6D0
LINE
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
CTLJSHJ
SOORCE
; CHECK IF ![HE START BUTTON HAS 1 BEEN PRESSED
AL,PORT1
LOOP1: IN
AND AL,START BIT
JNZ LOOP1
; SEND THE MICROWAVE POLSE OOT
START: MOV AL.PULSl
OOT PORTO,AL
; CHECK IF THE ULTRASOUND POLSE 1IAS ARRIVED '
; IF 'YES', THEN START THE CONVElIS I ON
AL,P0RT1
L00P2: IN
AND AL,ST CONVRT
JZ
LOOP2
;INITIALIZING THE A_TO_D
LOP: IN
AL.PORT1
AND AL,BUSY
JZ
LOP
AL,A TO D
IN
CONVST:OUT A_TO_D,AL
NOP
»
; CHECK IF 1fHE CONVERSION IS DONI
AL.PORT1
LOOP3: IN
AND AL.BUSY
L00E3
JZ
; IHPUT THE DATA AND STORE IN TBIi ARRAY
AL,A TO D
NEXT: IN
NOT AL
160
78
79
80
81
82
83
04
85
CO
C5
003E 268807
0041 43
0042 B000
0044 268807
0047 43
0048 49
0049 75E6
004B 4B
004C B061
004E E6B5
0050 5D
0051 IF
0052 CB
—
87
AP
00
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
ASSEMBLY COHPLETE, NO ERRORS FOUND
MOV ES:[BX],AL
INC BX
HOV AL,OH
MOV ES:[BX],AL
INC BX
; START THE NEXT CONVERSION
;
OUT AJOJ.AL
; CHECK IF ALL 25 CONVERSIONS HAVE BEEN FINISHED
;
IN AL,PORT1
;
AND AL,DONE
JZ CONVRT
DEC CX
JNZ CONVRT
;RESET A_TO_D
;
IN ALjAJOJ
; CLEAR ALL'THE FLIP-FLOPS AND THE COUNTER AT THE 2ND LEVEL
DEC BX
MOV AL.INIT3
OUT PORTO,AL
; RETURN TO THE MAIN PROGRAM
POP BP
POP DS
RET
INJATAJ ENDP
DATAJN ENDS
END
8086/87/88/186 MACRO ASSEMBLER
DISPLAY
iRHX 36 8086/87/88/186 MACRO ASSEMBLER V2.0 ASSEHBLY OF HODOLI DISPLAY
OBJECT MODULE PLACED IN :FDO:DSPY.OBJ
ASSEMBLER INVOKED BY: :LANG:ASH86 :FDO:DSPY.ASH
)C OBJ
...
00A0
OOAO
00A1
00A2
OOA3
00A4
00A5
00A6
00A6
00A7
00A8
OOA9
OOAA
OOAB
OOAC
OOAC
OOAE
OOAF
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOC
000D
...
...
LINE
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
SOORCE
SAME DISPLAY
; REGISTER LOCATION
PROGJATA SEGMENT PUBLIC
EQO OAOH
BASE
EQO BASE+OOH
X LO
EQU BASE+01H
X HI
EQO BASE+02H
Y LO
EQO BASE+03H
Y HI
EQU BASE+04H
DATA LO
EQO BASE+05H
DATA HI
EQU BASE+06H
STATUS
EQU BASE+06H
CONTROL 1
EQO BASE+07H
CONTROL 2
CONTROL 3 LO EQU BASE+08H
CONTROL 3 'HI EQO BASE+09H
CONTROL 4_LO EQU BASE+OAH
CONTROL 4 HI EQU BASE+OBH
EQU BASE+OCH
CRTC STATUS
EQU BASE+OCH
CRTC ADDR
EQU BASE+OEH
CRTC DATA
EQU BASE+OFH
VECTOR
EQO OOH
CRTC RO
EQO 01H
CRTC R1
EQU 02H
CRTC R2
EQU 03H
CRTCJ3
EQU 04H
CRTCJ4
EQU 05H
CRTC R5
EQU 06H
CRTC R6
EQU 07H
CRTC R7
EQU 08H
CRTCJ8
EQU 09H
CRTC R9
EQU OCH
CRTC R12
CRTC R13
EQO ODH
PROGJATA ENDS
PROGjCODE SEGHENT POBLIC
PUBLIC INITJHAGE
POBLIC CLEAR JSPY
POBLIC DSPYJHAGE
; INITIALIZE THE IMAGE BOARD
; CLEAR THE DISPLAY
; DISPLAY THE IMAGE IN 2D
163
0000 IE
0001 B8—
R
43
44
45
46
47
48
49
50
PDBLIC DSPYJRRAY
; DISPLAY AH OSE DIMENSIONAL ARRAY
PDBLIC DSPY_CODE
; DISPLAY TH
;
iNITJMAGE PROC FAR
POSH DS
ASSUME CS:PROG_CODE,DS:PROGJATA
MOV AX,PROG JATA jINITIALIZE DS
MOV DS,AX
8086/87/88/186 MACRO ASSEMBLER
DISPLAY
LINE
SOURCE
LOC OBJ
0006 55
0007 B020
0009 E6A6
000B E6A6
000D B03F
000F E6A7
0011 BOOF
0013 E6A8
0015 BOOO
0017 E6AA
0019 BOOO
001B E6AC
001D B04F
001F E6AE
0021 B001
0023 E6AC
0025 B040
0027 E6AE
0029 B002
002B E6AC
002D B045
002! B6AE
0031 B003
0033 E6AC
0035 B036
0037 E6AE
0039 B004
003B E6AC
003D B03F
003E E6AE
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
PUSH BP
; INITIALIZE THE IMAGE BOARD
MOV
OUT
OUT
MOV
OUT
MOV
OUT
MOV
OUT
MOV
OUT
MOV
ODT
MOV
OUT
MOV
OUT
MOV
OUT
MOV
OUT
MOV
ODT
HOY
OUT
MOV
OUT
MOV
ODT
AL.20H
CONTROLJ,AL
COHTROL_l,AL
AL.3FH
CONTROL_2,AL
AL,0FH
CONTROL_3_LO,AL
AL.OOH
CONTROL_4JiO,AL
AL,CRTC_R0
CRTC ADDR,AL
AL.4FH
CRTC_DATA,AL
AL.CRTCJl
CRTC ADDR.AL
AL.40H
CRTC DATA.AL
AL.CRTCJ2
CRTCJDDR.AL
AL.45H
CRTCJATA,AL
AL.CRTC R3
CRTCJDDR.AL
AL,36H
CRTCJATA,AL
AL, CRTCJ4
CRTC ADDR.AL
AL,3FH
CRTC DATA,AL
HOV AL.CRTC R5
OOT CRTC ADDR.AL
HOV AL,04H
ODT CRTCJATA.AL
HOV AL,CRTCJ6
OOT CRTCJDDR.AL
HOV AL.3CH
OOT CRTCJATA.AL
HOV AL.CRTCJ7
OOT CRTCJDDR.AL
HOV AL.3DH
OOT CRTCJATA.AL
HOV AL.CRTC R8
OUT CRTC ADDR, AL
HOV AL,07H
OOT CRTCJATA.AL
• HOV AL.CRTCJ9
OUT CRTCJDDR.AL
HOV AL.07H
ODT CRTC DATA.AL
HOV AL.CRTCJ12
ODT CRTCJDDR.AL
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
0041 B005
0043 E6AC
0045 B004
0047 E6AE
0049 B006
004B E6AC
004D B03C
004F E6AE
0051 B007
0053 E6AC
0055 B03D
0057 E6AE
0059 B008
005B E6AC
005D B007
005F E6AE
0061 B009
0063 E6AC
0065 B007
0067 E6AE
0069 BOOC
006B E6AC
8086/87/88/186 HA(;RO ASSEMBLER
DISPLAY
LOC OBJ
LIHE
SOURCE
006D BOOO
006F E6AE
0071 BOOD
0073 E6AC
0075 BOOO
0077 E6AE
0079 5D
007A IF
007B CB
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
HOV AL.OOH
ODT CRTC DATA.AL
HOV AL.CRTCJ13
OUT CRTC ADDR,AL
HOV AL.OOH
OUT CRTCJATA.AL
POP BP
POP DS
RET
1HITJHAGE EHDP
007C
007C IE
007D B8—
0080 8ED8
0082 55
R
1
1
CLEARJSPY PROC FAR
POSH DS
ASSUME CS:PROGJCODE.DS:PROGJATI
HOV AX.PROGJATA ;IHITIALIZE DS
HOV DS,AX
POSH BP
; CLEAR THE DISPLAY
!
0083
0085
0087
0089
008B
008C
008D
BOOO
E6A4
BOAO
E6A6
5D
IF
CB
008E
008E
008F
0090
0091
0092
0093
0094
5A
59
5B
5E
51
52
IE
0095
0098
009A
009B
009D
B8—
8ED8
55
8BEB
8EC6
009F
00A1
00A3
00A6
00A9
OOAB
BOOF
E6A8
BEOOOO
BF2E00
B507
B108
OOAD 46
OOAE 8BC6
OOBO E6A2
S
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
8086/87/88/186 HACRO ASSEMBLER
LOC OBJ
00B2
00B4
00B6
00B8
OOBA
OOBC
OOBE
8AC4
E6A3
8BDD
8BC7
E6A0
8AC4
E6A1
LINE
161
162
163
164
165
166
167
HOV
OOT
HOV
OOT
POP
POP
RET
CLEARJSPY ENDP
AL.OOH
DAT4J.0.AL
AL.OAOH
CONTROL 1.AL
BP
DS
I
; DISPLAY
DSPYJHAGE PROC FAR
POP DX
POP CX
POP BX
POP SI
POSH CX
POSH DX
POSH DS
ASSOHE CS:PROG CODE,DS:PFOG
HOV AX,PROG DATA ;INITIALIZE DS
HOV DS.AX
POSH BP
HOV BP.BX
HOV ES.SI
1
L00P1:
HOV
OOT
HOV
HOV
HOV
HOV
AL.OFH
CONTROL 3 LO.AL
SI.OOOOH
DI.002EH
CH.07H
CL,08fl
1
L00P2:
INC SI
HOV AX,SI
OOT Y LO.AL
DISPLAY
SOURCE
L00P3:
HOV
OOT
HOV
HOV
OOT
HOV
OUT
AL,AH
YJI.AL
BX,BP
AI.DI
X_LO,AL
AL.AH
XJI.AL
00C0
00C3
00C5
00C7
OOC9
OOCB
OOCC
268A07
E6A4
FECD
7504
B507
43
43
OOCD
OOCE
OODO
00D3
47
8BC7
3DD101
7EE3
00D5 BF2E00
00D8 FEC9
OODA 75D1
OODC
OODE
OOEO
00E3
8BEB
8BC6
3DDF01
7EC6
00E5 5D
00E6 IF
00E7 CB
00E8
00E8 58
00E9 5A
OOEA 5F
OOEB 5E
OOEC 5B
OOED 59
OOEE 8EC1
OOFO 268A2F
00F3 8BDF
OOF5 52
00F6 50
00F7 IE
00F8
OOFB
OOFD
OOFE
B8—
8ED8
55
8EC6
R
168
169
170
171
172
173
174
175
176
177
178
• 179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
8086/87/88/186 MACRO ASSEMBLER
166
)
}
MOV
OUT
DEC
JNZ
HOV
INC
INC
AL.BYTE PTR ES:[BX]
DATA LO.AL
CH
NEXT
CH.07H
BX
BX
INC
MOV
CMP
JLE
DI
AX.DI
AX.01D1H
L00P3
1
NEXT:
J
HOV DI.002EH
J
DEC CL
JNZ L00P2
I
HOV
HOV
CHP
JLE
BP,BX
AX,SI
AX,01DFH
L00P1
1
POP BP
POP DS
RET
DSPYJHAGE ENDP
)
DSPYJRRAY PROC FAR
POP AX
POP DX
POP DI
POP SI
POP BX
POP CX
HOV ES.CX
HOV CH.BYTE PTR ES:[BX]
HOV BX.DI
POSH DX
POSH AX
POSH DS
ASSOHE CS:PROGJCODE,DS:PS
HOV AX.PROGJATA ;INITIAL
HOV DS.AX
POSH BP
HOV ES.SI
DISPLAY
LOC OBJ
0100
0102
0104
0107
BOOF
E6A8
BEE001
BF2E00
OlOA B107
OlOC 8BC7
OlOE E6A0
0110 8AC4
0112 E6A1
0114 268B17
0117 23D2
0119 7410
OllB 4E
011C 8BC6
OllE E6A2
0120 8AC4
0122 E6A3
0124 BOOD
0126 E6A4
0128 4A
0129 75F0
012B 47
012C 8BC7
012E E6A0
0130 8AC4
0132 E6A1
0134 BEE001
0137 268B17
013A FEC9
013C 75D6
013E 43
013F 43
0140 FECD
0142 75C6
0144 5D
0145 IF
0146 CB
0147
0147
0148
0149
014A
014B
014C
014D
5A
59
5B
5E
51
52
IE
LINE
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
167
SOURCE
; DISPLAY TBE HISTOGRAH/EDGE PROFILE
HOV AL.OFH
OOT CONTROL 3 LO.AL
HOV SI,01E0H
HOV DI.002EH
L00P4:
L00P5:
LOP5:
HOV CL.07
HOV AX.DI
OOT XJIO.AL
HOV AL.AH
OOT X HI.AL
HOV DX.HORD PTR ES:[BX]
AND DX,DX
JZ L00P6
DEC SI
HOV AX,SI
OOT Y LO.AL
HOV AL.AH
OOT Y HI.AL
HOV AL.ODH
OOT DATA_LO,AL
; SET OP THE X-ADDRESS
; LOAD THE HISTOGRAH VALOES TO AX
; SET OP THE YJDD8ESS
nj?p IM
HY
l/uv
JNZ LOP5
INC DI
HOV AX.DI
OOT X LO.AL
HOV AL.AH
OOT X HI.AL
HOV SI,01E0H
HOV DX.HORD PTR ES:[BX]
DEC CL
JNZ LOOP5
INC BX
INC BX
DEC CH
JNZ LOOP4
POP BP
POP DS
RET
DSPYJRRAY ENDP
L00P6:
DSPYjCODE PROC FAR
POP DX
POP CX
POP BX
POP SI
POSH CX
POSH DX
POSH DS
ASSUHE CS:PROG CODE.DS:
; NEXT SCAN LINE
; LOAD THE HISTOGRAH VALUES TO AX
IJATA
014E
0151
0153
0154
B8—
8ED8
55
8EC6
R
266
267
268
269
270
HOV AX,PBOG_DATA ;IHITIALIZE DS
HOV DS,AX
POSH BP
HOV ES,SI
;
8086/87/88/186 HACRO ASSEHBLER
DISPLAY
LOC OBJ
LIHE
SOORCE
0156
0158
015A
015D
0160
0162
BOOF
E6A8
BEOOOO
BFE601
B507
B11E
0164
0165
0167
0169
016B
016D
016F
0171
0173
46
8BC6
E6A2
8AC4
E6A3
8BC7
E6A0
8AC4
E6A1
0175
0178
017A
017B
017D
017F
0181
0184
0186
0188
0189
268A07
E6A4
47
FECD
75EE
B507
BFE601
FEC9
75DC
4B
4B
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
018A 8BC6
018C 3DDF01
018F 7ED1
0191 5D
0192 IF
0193 CB
—
LOOP7:
' HOV
ODT
HOV
HOV
HOV
HOV
AL.OFH
CONTROL 3 LO.AL
SI,0000B
DI.01E6B
CH.07H
CL,1EH
»
LOOPS:
L00P9:
INC
HOV
OOT
HOV
OOT
HOV
OOT
HOV
OOT
SI
AX,SI
Y LO.AL
AL,AH
Y HI,AL
AX.DI
X LO.AL
AL,AH
XJI.AL
HOV
OOT
INC
DEC
JNZ
HOV
HOV
DEC
JNZ
DEC
DEC
AL.BYTE PTR ES:[BX]
DATA LO.AL
DI
CH
L00P9
CH.07H
DI,01E6H
CL
LOOP8
BX
BX
I
1
HOV AX,SI
CUP AX.01DFH
JLE LOOP7
1
POP BP
POP DS
RET
DSPYJODE ENDP
PROG CODE ENDS
309
310
ASSEMBLY COMPLETE, HO EBRORS FOUND
170
VITA
Karen H. Chan received her B.S. in electrical engineering from
Virginia Polytechnic And State University in 1979, her M.S. in
electrical engineering from the Ohio State University in 1981, and her
Ph. D. in bioengineering from University of Illinois at Chicago in
1988.
Her research interests include pattern recognition, image
processing and biomedical instrumentation.
HONORS AND AWARDS
Graduate School Fellowship, University of Illinois at the Medical
Center; September 1982- August 1983.
University Fellowship, University of Illinois at Chicago Circle;
September 1981 - June 1982.
Phi Kappa Phi Honor Scociety; 1979.
The Claude H. Brown Award for outstanding freshman in. mathematics;
1976.
SELECTED ABSTRACTS AND PUBLICATIONS
Lin, J., Chan, K. and Popovic, M.A.: Dual-frequency cardiopulmonary
rate monitor, in poster section, BEMS, June 1984.
Lin, J.C., and Chan, K.H.: Microwave thermoelastic tissue imagingsystem design, IEEE Trans. Microwave Theory Tech., 32:854-860,
1984.
Chan, K. and Lin, J.: Microprocessor-based cardiopulmonary rate
monitor, Medical & Biological Engineering and Computing, 41-44,
January 1987
Sychra, J.S., Pavel, D.G., Capek, V., Horowitz, A., and Chan, K.:
Synthetic NMR Images with increased information content and
decreased noise, presented at the 5th meeting of Society of
Magnetic Resonance Imaging, San Antonio, Texas, 1987.
PROFESSIONAL MEMBERSHIP
Member of the Computer Society, IEEE
Member of the Society of Biology, Medicine and Engineering, IEEE
Member of the Society of Women Engineers
Документ
Категория
Без категории
Просмотров
0
Размер файла
5 680 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа