Paper Details
Deep Learning-Based Glaucoma Detection via Convolutional Neural Networks
Authors
Ramesh Chouhan, Prof. Vikas Kalme
Abstract
Glaucoma is a serious condition that can lead to permanent blindness, if not caught and treated early. So, an automatic and precise system which can detect glaucoma is at a high need. Deep learning based methods have demonstrated great potential for medical image analysis in recent years, such as glaucoma detection. In this paper, we present a CNN-based system to detect glaucoma using a publicly available dataset. Since the proposed system is supposed to detect glaucoma in digital retinal fundus images, automatically and with high accuracy. These results show the feasibility and clinical applicability of the method. This study used dropout method to enhance umbrella of accuracy glaucoma detection. Experiments on the SCES and ORIGA datasets have proven that our method is effective. The method proposed achieved an accuracy of 99.12% on the ORIGA dataset and an accuracy of 99.37% on the SCES dataset. Using state-of-the-art techniques, we found that ORIGA was 86% accurate, and SCES was 91%.
Keywords
Image Processing, Glaucoma Diagnosis, Image Registration, Fusion, Segmentation, Statistical Measures, Morphology, Classification, Pattern Matching.
Citation
Deep Learning-Based Glaucoma Detection via Convolutional Neural Networks. Ramesh Chouhan, Prof. Vikas Kalme. 2024. IJIRCT, Volume 10, Issue 5. Pages 1-11. https://www.ijirct.org/viewPaper.php?paperId=2409031