Dynamic region proposal networks for semantic segmentation in automated glaucoma screening

TL;DR

This paper introduces two novel neural network approaches for segmenting optic disc and cup in fundus images to aid glaucoma screening, achieving comparable accuracy with fewer parameters by using dynamic cropping and end-to-end training.

Contribution

The paper presents Parameter-Shared Branched Network and Weak Region of Interest Model-based segmentation, which are trained end-to-end and utilize dynamic cropping, reducing model complexity while maintaining high accuracy.

Findings

Achieves Dice scores of 0.96/0.89 on Drishti-GS1 dataset.

Uses significantly fewer parameters than state-of-the-art methods.

Demonstrates effective segmentation performance on public datasets.

Abstract

Screening for the diagnosis of glaucoma through a fundus image can be determined by the optic cup to disc diameter ratio (CDR), which requires the segmentation of the cup and disc regions. In this paper, we propose two novel approaches, namely Parameter-Shared Branched Network (PSBN) andWeak Region of Interest Model-based segmentation (WRoIM) to identify disc and cup boundaries. Unlike the previous approaches, the proposed methods are trained end-to-end through a single neural network architecture and use dynamic cropping instead of manual or traditional computer vision-based cropping. We are able to achieve similar performance as that of state-of-the-art approaches with less number of network parameters. Our experiments include comparison with different best known methods on publicly available Drishti-GS1 and RIM-ONE v3 datasets. With $7.8 \times 10^6$ parameters our approach achieves a Dice score of 0.96/0.89 for disc/cup segmentation on Drishti-GS1 data whereas the existing state-of-the-art approach uses $19.8\times 10^6$ parameters to achieve a dice score of 0.97/0.89.