Glaucoma detection using novel perceptron based convolutional multi-layer neural network classification

Glaucoma in retina is considered as a significant cause of irreparable vision loss. For the automated glaucoma detection on fundus images, several approaches have been recently developed. However, the extraction of optic cup (OC) boundary considered as critical work due to the blood vessels interweavement. To accurately detect glaucoma and correspondingly the optic cup OC and optic disk (OD) boundary, the Perceptron based Convolutional Multi-Layer Neural network classification with Weighted Least Square Fit holistic feature extraction has been proposed. In this study, the Glaucoma classification has been performed using Perceptron based Convolutional Multi-Layer Neural network. The optic disc and optic cup boundary is detected by entropy based estimation. After the optic disc and cup boundary are segmented, the disc ratio and holistic local features are extracted by Weighted Least square fit. The Perceptron based Convolutional Multi-Layer Neural network classification is performed for various datasets such as Drishti-GS and Rim-ONE dataset to accurately classify the Glaucoma in retinal images. The proposed method is evaluated for the OC and OD segmentation, accurate glaucoma detection in terms of accuracy, sensitivity, specificity, dice and overlap, and Jaccard parameters.