DETECTION OF DIABETIC MACULAR EDEMA IN OPTICAL COHERENCE TOMOGRAPHY SCANS USING PATCH BASED DEEP LEARNING

We propose a two step framework to automatically classify an OCT scan as indicative of Diabetic Macular Edema (DME) by detecting abnormal pathologies in OCT frames. The first step involves detection of candidate patches for fluid filled regions and hard exudates using image processing techniques. The second step is to predict a label for these candidate patches using deep convolutional neural network. In the final collation step, we aggregate the confidences of the CNN models and use a rule based method to predict the presence of DME.