Two hybrid CNN Algorithms with An Android Application for Detection of GLAUCOMA

Abstract

The human eye is one of the body's most important organs. The eye is constantly important in our daily lives; without eyes, the world would be dark and doing daily tasks would be extremely difficult. In the sense that without sight, it would be extremely difficult for anyone to perform any task. The loss of vision/sight in the human eyes can be caused by a variety of factors. As a result, blindness in the human eyes must be prevented, as the most valuable human organ is solely responsible for vision. Different forms of diseases that occur in the eyes as a result of numerous circumstances are one of the causes of blindness and visual loss in the eyes. One such sickness is one that develops as a result of Convolutional Neural Network (CNN) is being offered as a way to diagnose glaucoma using fundus pictures of the eyes. In the proposed algorithm we use, k-means algorithm for segmentation, GLCM for feature extraction and classify using Multi-SVM (Support Vector Machine) as first hybrid algorithm & we use Otsu thresholding method for segmenting then, HOG (FE) feature extraction techniques is used & classification based on Knn algorithm as second hybrid algorithm & implement the same for creating an android mobile application.