вход по аккаунту


978-3-319-68195-5 82

код для вставкиСкачать
LED-based System for the Quantification
of Oxygen in Skin: Proof of Concept
Pérez Sandra1(B) , Tapia Pedro2 , Galeano July3 , Zarzycki Artur5 ,
Garzón Johnson4 , and Marzani Franck6
Grupo de Investigación e Innovación Biomédica- Instituto Tecnológico
Metropolitano, Grupo de Investigaciones en Bioingeniera,
Universidad Pontificia Bolivariana, Circ. 1 73-76, Medellı́n, Colombia
Grupo de Investigaciones en Bioingeniera, Universidad Pontificia Bolivariana,
Circ. 1 73-76, Medellı́n, Colombia
Grupo de Investigacion en Materiales Avanzados y Energia MatyEr,
Linea de Investigacion Biomateriales y Electromedicina,
Calle 54 A 30-01, Medellı́n, Colombia
UPB, Grupo de Óptica y Espectroscopı́a, Circular 1 70-01, Medellı́n, Colombia
Grupo de Investigación en Automática y Electrónica,
Linea Sistemas de Control y Robótica, Calle 54 A 30-01, Medellı́n, Colombia
Laboratoire Le2i, UFR Sciences et Techniques, Universit de Bourgogne,
B.P. 47870, 21078 Dijon Cedex, France
Abstract. Imaging technologies have the potential to supplement conventional diagnosis and follow up of ulcers, by providing detailed information regarding skin components imperceptible to visual inspection.
Clinical investigations have shown that several factors including reduced
oxygen delivery and disturbed metabolism can impair the wound healing process, for that reason is desirable to asses the distribution of tissue oxygenation around a lesion. In this sense, this work describes the
development and preliminary tests of a LED-based multispectral imaging
system to measure changes in the oxygen saturation.
Keywords: Multispectral imaging
· Oxygen saturation · Skin ulcers
Skin ulcers are a major public health concern, which have become one of the
most frequent causes of consultation in primary health-care units in tropical and
subtropical countries. In developed countries, chronic wounds affect more than
c Springer International Publishing AG 2018
J.M.R.S. Tavares and R.M. Natal Jorge (eds.), VipIMAGE 2017,
Lecture Notes in Computational Vision and Biomechanics 27,
DOI 10.1007/978-3-319-68195-5 82
LED-based System for the Quantification of Oxygen in Skin
1% of the population and are most commonly due to circulatory or metabolic
disorders and their incidence is expected to follow those observed for diabetes
and obesity [3]. Oxygen is a prerequisite for successful wound healing due to the
increased demand for reparative processes such as cell proliferation, bacterial
defence, angiogenesis and collagen synthesis. Even though the role of oxygen in
wound healing is not yet completely understood, many experimental and clinical
observations have shown wound healing to be impaired under hypoxia [11].
Multispectral imaging systems allow obtaining information outside the
human perception range, which may include analysis in ultra violet, visible,
infrared bands, X-ray or other bands of the electromagnetic spectrum [4]. In
medicine these images are an effective tool, mainly for cancer diagnosis [2,9].
Also, multispectral imaging is investigated for diagnosis and follow up of skin
pathologies [6]. There are some studies focused on the area of oxygen quantification, employing complexes models, no visible light or expensive hardware
This paper presents a LED-based system to acquire and process multispectral
images in order to measure variations in oxygen saturation in skin. Also, the
proof of concept was performed acquiring images of 30 index fingers of volunteers
and the results were compared with a comercial oximeter.
Materials and Methods
Experimental Setup and Image Acquisition
The light source selected was an O’ring LED Neo Pixel with nine wavelengths
programmed by the open-source electronics platform Arduino. LED’s illumination system was characterized using an Ocean Optics spectrometer (HR4000
UC-UV-NIR), in order to obtain the central wavelengths. To collect the reflected
light, a vision system composed by a high sensitivity CCD camera (DCC-3240N
CMOS Thorlabs) and a 50 mm lens (C-Series VIS-NIR Edmund Optics) was
implemented. Finally, the depth of field, working distance and focal length of
the imaging system were adjusted.
Multispectral images were acquired from the index finger of 30 human volunteers in conditions of normal oxygenation and applying a tourniquet, in order
to obtain lower values of oxygen saturation (SpO2 ) [7]. Fingers were selected
as area under test because it was necessary monitoring, with an oximeter, the
SpO2 before and after reducing blood flow. Implementing this methodology, one
image for each wavelength was obtained for high and low SpO2 conditions.
Image Processing
In order to convert the raw intensity into reflectance, reference and dark images
were captured before acquiring fingers images. The reference image was obtained
with a white diffuse reflectance standard, that reflects 99% of light, placed in the
P. Sandra et al.
scene; and the dark current is measured by keeping the camera shutter closed.
The raw data were then adjusted employing the Eq. 1 [8].
Iref =
Iraw − Idark
Iwhite − Idark
where, Iref is the tissue multispectral image after white balance (representing
tissue reflectance spectra), Iraw is the pixel intensity from the tissue multispectral image without compensation, Idark corresponds to the camera dark current,
and Iwhite is the pixel intensity given by the white standard image.
Data Analysis
The mean values of reflectance, calculated from ten pixels of each finger image,
were established in relation to the factors oxygen saturation level (two groups:
high or low according to oximeter measurements) and the nine wavelengths generated by the light source.
Multifactorial analysis of variance (ANOVA) was used to compare the measured reflectance with the factors and evaluating if they have a statistically
significant effect on reflectance. Then, additional tests was applied to find if the
two groups are actually different.
Experimental Setup and Image Acquisition
Central wavelengths obtained from spectral characterization were 464 nm,
490 nm, 525 nm, 550 nm, 610 nm, 619 nm, 620 nm and 630 nm. Camera exposure
time was manually adjusted based on white calibration: 5.1 ms, 7.3 ms, 4.7 ms,
4.5 ms, 4.4 ms, 4.4 ms, 4.4 ms, 4.4 ms and 4.4 ms respectively for each wavelength.
The schematic diagram of the multispectral imaging system is presented in Fig. 1.
Images obtained from one of the volunteers index finger at 464 nm, 525 nm
and 630 nm; before and after applying the tourniquet are shown in Fig. 2.
Image Processing
Figure 3 presents the obtained curves from plotting wavelength versus the Iref of
index fingers, from 3 of the 20 volunteers, lighted with the multispectral system.
Calculations were performed in 10 different pixels of an specific region of interest.
Data Analysis
Analysis was performed using only images acquired illuminating the volunteers
fingers with 630 nm, 525 nm and 464 nm (RGB) wavelengths, because the others
are formed by mixing spectral light in varying combinations.
LED-based System for the Quantification of Oxygen in Skin
Fig. 1. Configuration of the multispectral system based on LED.
Fig. 2. Index fingers lighted with (a) 464 nm, (b) 525 nm, (c) 630 nm
P. Sandra et al.
Fig. 3. Comparison of high and low SpO2 of tree different volunteers. Blue points
represents the measurements performed under the effect of the tourniquet (low SpO2
levels) and the red ones are in normal conditions of oxygenation (high SpO2 levels)
Table 1. Analysis of variance for reflectance
Sum of squares Df
Mean square F-ratio P-value
Main effects
SpO2 level
Wavelength 0.9669
1 0.00492
2 0.4834
392,37 0.0000
2 116
Analysis of variance for reflectance: Table 1 presents the variability decomposition of reflectance (Iref ) into contributions due to oxygenation and wavelength.
Since Type III sums of squares have been chosen, the contribution of each factor
was measured having removed the effects of all other factors. Being that, both
P-values were less than 0,05, these factors have a statistically significant effect
on Reflectance at the 95,0% confidence level.
Multiple range tests for reflectance by SpO2 levels: Fisher’s least significant
difference (LSD) procedure used to determine which means are significantly dif-
LED-based System for the Quantification of Oxygen in Skin
Table 2. Multiple range tests for reflectance by SpO2 level
SpO2 Level
LS Mean
LS Sigma
+/- Limits
denotes a statistically significant difference.
ferent from which others were applied (Table 2). The bottom half of the Table 2
shows the estimated difference between each pair of means. An asterisk is placed
next to 1 pair, indicating that this pair shows a statistically significant difference
at the 95.0% confidence level. At the top of the Table 2, two homogenous groups
are identified using columns of X’s. Within each column, the levels containing
X’s form a group of means within which there are no statistically significant
Means plot: Fig. 4 shows the mean Reflectance for each level of the factors. It
also shows the standard error of each mean, which is a measure of its sampling
variability. The rightmost two columns show 95,0% confidence intervals for each
of the means.
Fig. 4. Means and 95% LSD intervals
P. Sandra et al.
A Led-based system to acquire multispectral images were designed and implemented. The results obtained from comparing reflectance at different wavelengths and the oxygen saturation level, indicates that with this system is feasible
distinguishing between low and high values of SpO2 . For further works, an standard to compare the ulcer borders SpO2 values and the obtained with the system
is required.
Acknowledgements. The authors would like to acknowledge to Regional Program
Stic-AMSud for the support in the strengthening of the academic network.
1. Anand, S., Sujatha, N., Narayanamurthy, V.B., Seshadri, V., Poddar, R.: Diffuse
reflectance spectroscopy for monitoring diabetic foot ulcer – a pilot study. Opt.
Lasers Eng. 53, 1–5 (2014)
2. Galeano, J., Jolivot, R., Benezeth, Y., Marzani, F., Emile, J., Lamarque, D.: Analysis of Multispectral Images of Excised Colon Tissue Samples Based on Genetic
Algorithms (2014). De Bourgogne, Dijon Cedex, Phahonyothin Rd, and Boulogne
3. Guenin-Macé, L., Oldenburg, R., Chrétien, F., Demangel, C.: Pathogenesis of skin
ulcers: lessons from the Mycobacterium ulcerans and Leishmania spp. pathogens.
Cell. Mol. Life Sci. 71(13), 2443–2450 (2014)
4. Hardeberg, J., Schmitt, F., Brettel, H., Crettez, J., Maı̂tre, H.: Multispectral image
acquisition and simulation of illuminant changes. Colour Imag.: Vis. Technol. 2–19
5. Jeffcoate, W.J., Clark, D.J., Savic, N., Rodmell, P.I., Hinchliffe, R.J., Musgrove,
A., Game, F.L.: Use of HSI to measure oxygen saturation in the lower limb and
its correlation with healing of foot ulcers in diabetes. Diabet. Med. 32(6), 798–802
6. Jolivot, R.: Development of an imaging system dedicated to the acquisition, analysis and multispectral characterisation of skin lesions. Ph.D. thesis (2012)
7. Li, J., Dunmire, B., Beach, K.W., Leotta, D.F.: A reflectance model for non-contact
mapping of venous oxygen saturation using a CCD camera. Opt. Commun. 308,
78–84 (2013)
8. Lu, G., Fei, B.: Medical hyperspectral imaging: a review
9. Nouri, D., Lucas, Y., Treuillet, S., Jolivot, R., Marzani, F.: Colour and multispectral imaging for wound healing evaluation in the context of a comparative
preclinical study (2013)
10. Saito, T., Yamaguchi, H.: Optical imaging of hemoglobin oxygen saturation using
a small number of spectral images for endoscopic application. J. Biomed. Opt.
20(12), 126011–1–8 (2015)
11. Schreml, S., Szeimies, R.M., Prantl, L., Karrer, S., Landthaler, M., Babilas, P.:
Oxygen in acute and chronic wound healing. Br. J. Dermatol. 163(2), 257–268
12. Zuzak, K.J., Gladwin, M.T., Cannon III, R.O., Levin, I.W.: Imaging hemoglobin
oxygen saturation in sickle cell disease patients using noninvasive visible reflectance
hyperspectral techniques: effects of nitric oxide. 0510, 1183–1189 (2003)
Без категории
Размер файла
1 265 Кб
978, 68195, 319
Пожаловаться на содержимое документа