Research On Correlation Between Iris Features And Diseases

Studies have shown that human health can be reflected on some areas of the person’s faceand hands, but there is no specific data confirm that there is a certain contact between thehuman body health with the changes on iris color, texture, density, shape aspects on iridologyfield. In this paper, according to the basic theoretical knowledge of iridology, do research ofthe correlation between geometrical deformation of the pupil outline and stomach disease, theresearch is a case in point to explore the correlation between the iris features and disease. Thisarticle discusses respectively the iris image acquisition, the establishment of iris image library,pupil localization, pupil contour extraction, geometric deformation features extraction and thecorrelation between these features and stomach diseases.According to the iris image characteristics, this paper proposes a pupil localizationmethod. Firstly, color image is converted to R channel grayscale image to minimize the impactof the floodlight on pupil boundary. Using gamma transform to adjust the contrast of the pupiland the iris, then according to the gray minimum principle to position four points on the edgeof the pupil, using these points to realize the pupil’s primary location, according to the pupiledge gradient maximum principle to achieve precise localization of the pupil. At the end, use100iris images to do a localization experiment, the localization success rate is94%.In addition, a pupil’s contour extraction method of anti-light spot interference is given inthe paper. Firstly, extend the pupil circular region into a rectangular area and binary-convertthe area. Find all light spot regions and fill them by the information around them. To extractthe pupil contour and repair all the contour lines that may be influenced by some light spots.At the end, an experiment of contour extraction was done and rate for correcting extractioncontour is for200images.Finally, according to the pupil contour, three pupil geometric features are extracted out,including compactness, discrete degree and roughness, use self-built iris medical library of todo a statistical experiment on every feature and a multi-feature analysis experiment based onrough sets and fuzzy reasoning, two experimental results are given.