Precise Localization of Optic Disc Region for Accurate Glaucoma Diagnosis Using Deep Reinforcement Learning

Glaucoma is a leading cause of blindness worldwide, and early detection is crucial for effective management. In this research will propose a novel methodology for the detection of glaucoma in retinal fundus images using deep reinforcement learning and other modern AI-based algorithms. The proposed method includes This program is built using deep reinforcement learning for active object localization, as well as machine learning techniques. For input picture validation and active object localization for Region of Interest (ROI) detection and feature extracts using CNN, a deep reinforcement learning model with a deep Q network is used on the basis of MDP. The model is tailored to a specific class and allows an agent to focus on areas with the potential to identify the exact location of pressure variation in fundus images happening in the Optic Disc (OD) region called the glaucomatous Region. The agent is instructed to fine-tune a bounding box by means of straightforward transformational actions in order to precisely determine the location of the Optic Disc. The localization agent being discussed is developed using deep reinforcement learning techniques and its performance is measured using the Kaggle fundus images dataset.