CLASSIFYING THE VULNERABILITIES IN HUMAN EYE USING Q-LEARNING BASED DENSENET ALGORITHM

Abstract

An early diagnosis and effective treatment of ocular diseases relies on the detection and classification of vulnerabilities in the human eye. Many times, high accuracy and recall retinal image weakness classification is beyond the capabilities of present methods. This paper provides a strong framework based on a Q-Learning (QL) based DenseNet model to improve the efficiency, reduce false positives, and improve the accuracy of vulnerability identification in retinal images. With higher precision and recall rates than with traditional methods, the developed QL based DenseNet system classifies retinal image vulnerabilities. The results indicate 10% better categorization, fifteen percent fewer false positives, and 20% higher processing efficiency.