Automatic Detection of Eye Cataracts and Disease Classification Using Hybrid Techniques

Main Article Content

B Ramesh Kumar, Shimna M. P

Abstract

Medical image analysis is the most demanding technology now days. The proposed system performs automatic cataract from the digital eye image and retinal fundus images. The proposed system has developed a new technique with a set of algorithms. Currently, methods available for cataract detection are only based on certain features, but medical images may have heterogeneous feature set; the main motive behind this work is to develop less iterative, effective multi-feature based eye image analysis and cataract detection from the color images of eye and retinal fundus images. An algorithm is proposed for Cataract Screening based on the retinal features, veins, blood vessels, and artery. These features are used to analyze and classify the eye into specific class. To achieve this set of algorithms have proposed. The proposed system performs the pre-processing step initially, and then the feature selection from the preprocessed images is then initially classified using the Kernel Hyper Support Vector Machine (KHSVM). The results from the KHSVM, the effective features are applied into the modified genetic algorithm named as IIGA (Iterative Intensity Genetic Algorithm); this performs a new type of gene selection from the KHSVM features. Instead of selecting the random features, the proposed system gets the features from the KHSVM result. The proposed system achieves better results than the existing works. The proposed system is implemented in Matlab tool with several eye images. The experimented result shows the proposed system achieved better detection than the existing techniques.

Article Details

How to Cite
, B. R. K. S. M. P. (2017). Automatic Detection of Eye Cataracts and Disease Classification Using Hybrid Techniques. International Journal on Future Revolution in Computer Science &Amp; Communication Engineering, 3(12), 304–308. Retrieved from http://www.ijfrcsce.org/index.php/ijfrcsce/article/view/414
Section
Articles