Accurate disease detection quantification of iris based retinal images using random implication image classifier technique

Abstract In a recent analysis, eye disease is complicated to predict the affected area by using fundus imaging. The eye disease like lesion, this lesion affected retinal image is due to unevenness illumination, contrast is low, and blurring image is analyzed, when this image is analysis not found the affected region of lesion affected retinal image in the conventional system. The conventional system is having some loss like accuracy less, and it takes more time to analyze the image, so the proposed technique is developed to diagnose eye retinal lesion disease to improve the image quality and increase the accuracy. The proposed Random Implication Image Classifier Technique (RIICT) method is developed in this system to classify the image in various color spaces can achieve excellent accuracy in different image analyzing. Preprocessing of the retinal image uses a median filter for noise removal and enhancement; this preprocessing process is covert the color image to grayscale image. In segmentation, the Discrete Image Clustering Technique (DICT) algorithms that are selected and arranged to produce seeds are used to separate the image foreground and background of the input image to find the affected region based on the similarity easily. The proposed Random Implication Image Classifier Technique (RIICT) algorithm is used to classifying the lesion results of this system. This proposed system detects the disease like cotton wool spot, lesion quickly, and classifies the development region in various iris images to handle in this system. Finally, the RIICT system is given a better accuracy result is 96.7%.