Detecting jaundice by using digital image processing
J. Castro-Ramos
1
, C. Toxqui-Quitl
2
, F. Villa Manriquez
1
, E.Orozco-Guillen
3
, A.
Padilla-Vivanco
2
,
4
JJ. Sánchez-Escobar
1
Instituto Nacional de Astrofísica, Óptica y Electrónica;
2
Univ. Politécnica de
Tulancingo;
3
Univ. Politécnica de Sinaloa,
4
Centro de Enseñanza Técnica Industrial
(México).
ABSTRACT
When strong Jaundice is presented, babies or adults should be subject to clinical exam like “serum
bilirubin” which can cause traumas in patients. Often jaundice is presented in liver disease such as
hepatitis or liver cancer. In order to avoid additional traumas we propose to detect jaundice (icterus) in
newborns or adults by using a not pain method. By acquiring digital images in color, in palm, soles and
forehead, we analyze RGB attributes and diffuse reflectance spectra as the parameter to characterize
patients with either jaundice or not, and we correlate that parameters with the level of bilirubin. By
applying support vector machine we distinguish between healthy and sick patients.
Keywords: Digital Image Processing, Diffuse Reflectance, Bilirubin, Support Vector Machine.
1. INTRODUCTION
Many methods have been developed to determine the level of bilirubin in newborns or adults both
invasive and non invasive, Penhaker[1] designed an electronic instrument based on transmitted light
through the skin. The skin photo-diagnostic handle 450 nm - 575 nm monochromatic light. Buttitta et al
[2] develop a non invasive bilirubin monitor which uses two wavelengths of light radiation directed to an
infant arterial system. The reflectivity or backscatter of the light from the infant`s bloodstream from the
light sources is detected an measured to determine the bilirubin level of that arterial system of the infant.
Alla Suresh et al [3] develop a method to isolate the intravascular and extra vascular bilirubin starting
with the diffuse reflectance spectrum. A nonlinear optimization algorithm was adopted to extract the
optical properties including bilirubin concentration from the skin reflectance spectrum. A support vector
machine (SVM) is a method for classifying multivariate data. The possibility of using SVM for develop
any diagnostic algorithms is also attracting attention. While Palmer et al.[4] used a linear SVM classifier
for classifying auto fluorescence and diffuse reflectance spectra of breast tissues in vitro, Lin et al.[5]
classified in vivo auto fluorescence spectra from tissues by using both the linear and the nonlinear SVM
classifier with RBF kernel. In the reports of both groups, the tissue spectra were dimensionally reduced
by applying linear PCA algorithms prior to using the SVM approach for classification. Lin et al.[5]
showed that the classification performance of an SVM classifier trained on the full spectral data was
comparable to that obtained with the classifier trained on the diagnostically relevant principal components
only. Their combined PCA-SVM approach was reported to have reduced computational complexity. In
this paper we develop a novel idea by taking an image we get the RGB attributes and by using a
spectrometer we get the diffuse reflectance spectra from patients with low and high level of bilirubin we
Three-Dimensional and Multidimensional Microscopy: Image Acquisition and Processing XXI,
edited by Thomas G. Brown, Carol J. Cogswell, Tony Wilson, Proc. of SPIE Vol. 8949, 89491U
© 2014 SPIE · CCC code: 1605-7422/14/$18 · doi: 10.1117/12.2041354
Proc. of SPIE Vol. 8949 89491U-1
Downloaded From: http://proceedings.spiedigitallibrary.org/ on 01/22/2015 Terms of Use: http://spiedl.org/terms