A Feature Extraction and Machine Learning Based Approach for Glaucoma Detection from Fundus Images

A chronic eye condition known as glaucoma causes progressive damage to the optic nerve and astrocytes as a result of excessive intraocular pressure (IOP). It is the second leading global contributor to vision loss and blindness. Stopping the disease’s progression towards total eyesight loss depends on early diagnosis. The diagnosis of glaucoma strongly depends on the expertise of a glaucoma expert ophthalmologist due to the complicated and varied disease pathology. To protect a patient’s eyesight, it is critical to identify glaucoma in its early stages. The ability to design algorithms that can automatically and accurately identify glaucoma from fundus images has been made possible by recent advancements in image processing and related machine learning techniques. This work presents a fundus image enhancement, segmentation, and denoising approach. Additional statistical measurements of the image have been made that might aid in the identification of glaucoma. When compared to current techniques, the suggested approach performs better than baseline procedures in terms of classification accuracy.