Computer Aided System For Early Detection Of Nuclear Cataract Using Circle Hough Transform

Abstract

Survey done by world health organization for probing causes of blindness indicates that cataract is major cause of blindness. Even age related cataract is most commonly observed, it is serious cause to think because of its appearance in minors and children for both eyes. Detecting cataract at earlier stage is challenge as it has less affecting vision. The three most general types of cataract are nuclear cataract, cortical cataract and post subcapsular cataract. Slit lamp observation with lens opacity classification system (LOCS-III) is used for detection and medical diagnosis by ophthalmologists. Lens replacement surgery is most common treatment suggested on cataract for correcting vision. Literature survey indicates towards the success and correctness of computer added detection and grading is function of correctness of lens localization from cataract eye image. The work presented in this paper uses slit lamp images from ophthalmologist at eye hospital with computer aided image processing to detect cataract at earlier stage. The challenge of detection of cataract at earlier stage is attended in steps like lens detection, lens segmentation, feature extraction and categorization. The overall accuracy is enhanced by use of Hough circle detection transform for lens detection and support vector machine for categorization. The detection and categorization is performed using statistical feature extraction with prior trained support vector machine.