Dynamic region proposal networks for semantic segmentation in automated glaucoma screening
This work addresses the need for efficient automated glaucoma screening tools for medical professionals, but it is incremental as it improves parameter efficiency without a major breakthrough in performance.
The paper tackles the problem of segmenting optic cup and disc regions in fundus images for glaucoma screening by proposing two novel methods, PSBN and WRoIM, which achieve similar performance to state-of-the-art approaches with fewer parameters, such as a Dice score of 0.96/0.89 for disc/cup segmentation on Drishti-GS1 data using 7.8 million parameters compared to 19.8 million parameters in existing methods.
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.