Convolutional Prompting for Broad-Domain Retinal Vessel Segmentation

TL;DR

This paper introduces Dual Convolutional Prompting (DCP), a plug-in module that enables a unified retinal vessel segmentation model to perform well across diverse image domains without altering the base network structure.

Contribution

The paper proposes DCP, a novel domain-specific feature extraction method that adapts existing segmentation networks for broad-domain retinal vessel segmentation.

Findings

DCP outperforms baseline methods on broad-domain datasets.

The unified model effectively handles varied imaging modalities.

Extensive experiments validate the effectiveness of DCP.

Abstract

Previous research on retinal vessel segmentation is targeted at a specific image domain, mostly color fundus photography (CFP). In this paper we make a brave attempt to attack a more challenging task of broad-domain retinal vessel segmentation (BD-RVS), which is to develop a unified model applicable to varied domains including CFP, SLO, UWF, OCTA and FFA. To that end, we propose Dual Convoltuional Prompting (DCP) that learns to extract domain-specific features by localized prompting along both position and channel dimensions. DCP is designed as a plug-in module that can effectively turn a R2AU-Net based vessel segmentation network to a unified model, yet without the need of modifying its network structure. For evaluation we build a broad-domain set using five public domain-specific datasets including ROSSA, FIVES, IOSTAR, PRIME-FP20 and VAMPIRE. In order to benchmark BD-RVS on the broad-domain dataset, we re-purpose a number of existing methods originally developed in other contexts, producing eight baseline methods in total. Extensive experiments show the the proposed method compares favorably against the baselines for BD-RVS.