Review of Machine Learning techniques for glaucoma detection and prediction

Glaucoma is a silent thief of sight. Detecting glaucoma at early stages is almost impossible and presently there is no cure of glaucoma at later stages. Different automated glaucoma detection systems were thoroughly analyzed in this study. A detailed literature survey of preprocessing, feature extraction, feature selection, Machine Learning (ML) techniques and data sets used for testing and training purpose was conducted. Automated prediction of glaucoma is very important and unfortunately a little work has been done in this regard and minimum accuracy has been achieved. However automated detection of glaucoma at latter stage is at a mature level and most of the ML techniques are able to detect 85% of glaucoma cases accurately. Optical Coherence Tomography (OCT) can be used effectively for prediction of glaucoma.