Abstract—There has been considerable progress in
computer vision, artificial neural network and pattern
recognition in the last two decades, and there has also much
progress in medical imaging technology in recent years.
Although images in digital form can be processed by basic
image processing techniques, effective use of computer vision
can provide much useful information for diagnosis and
treatment. In this paper we integrate computer vision and
iridology practice for the detection of diabetes. Using iridology
iris image is evaluated by detecting the presence of broken
tissues and change in color pattern. According to iridology the
abnormality in an iris of the human eye represent the
abnormality of the corresponding organ conferred by the iris
chart. In this research we examine pancreas organ which is at
position 01:45 – 02:15 for the right eye and 07:15-7:45 for the
left eye according to Dr. Jensen iris chart. We applied two
methods to reach our conclusion, visual inspection method and
color coding method. The artificial neural network is used for
training and classification purpose. The entire process is
showing a high accuracy detection of abnormality of pancreas
organ which led to diabetes. The final result is compared with
the insulin normality test for verification.
Index Terms—Computer vision, diabetic, feature extraction, iris, iridodiagnosis.
The authors are with Tianjin University, Tianjin, China (e-mail: firstname.lastname@example.org).
Cite: Jamal Firmat Banzi and Zhaojun Xue, "An Automated Tool for Non-contact, Real Time Early Detection of Diabetes by Computer Vision," International Journal of Machine Learning and Computing vol. 5, no. 3, pp. 225-229, 2015.