Screening of Glaucoma disease from retinal vessel images using semantic segmentation

A timely diagnosis of Glaucoma has crucial importance in preventing blindness. As this disease exists in the immediate vicinity of the optical disk (OD), its precise localization and segmentation are critical in its accurate diagnosis. OD consists of two parts, namely: neuroretinal and optic cup (OC). In the proposed work, the problem of OD and OC segmentation is modeled as a semantic pixel-wise labeling problem, thus bridging the gap between medical image segmentation and semantic segmentation. The proposed method eliminates the need for pre- and post-processing steps. The proposed method is evaluated for the segmentation of OD and OC on Drishti and Rim-one datasets. The offered low computational and resource requirements along with the observed state-of-the-art accuracy of the proposed method support its implementation in the real-time automatic screening of the Glaucoma disease.

Automated summary

The study presents an effective method for segmenting the optic disk (OD) and optic cup (OC) in glaucoma by semantic pixel-wise labeling, without the need for pre- and post-processing steps. Evaluated on different datasets, the method shows low computational and resource requirements, supporting its implementation in real-time screening for glaucoma disease.