The Unreasonable Effectiveness of Encoder-Decoder Networks for Retinal Vessel Segmentation

TL;DR

This paper introduces VLight, an encoder-decoder network for retinal vessel segmentation that achieves state-of-the-art accuracy, generalizes well across datasets, and is efficient with fewer than 0.8 million parameters.

Contribution

The paper presents a simple, fully-convolutional encoder-decoder framework that leverages multi-scale patch extraction for improved retinal vessel segmentation.

Findings

Achieves state-of-the-art results on three fundus datasets.

Demonstrates robustness and generalization across datasets.

Maintains high accuracy with fewer than 0.8M parameters.

Abstract

We propose an encoder-decoder framework for the segmentation of blood vessels in retinal images that relies on the extraction of large-scale patches at multiple image-scales during training. Experiments on three fundus image datasets demonstrate that this approach achieves state-of-the-art results and can be implemented using a simple and efficient fully-convolutional network with a parameter count of less than 0.8M. Furthermore, we show that this framework - called VLight - avoids overfitting to specific training images and generalizes well across different datasets, which makes it highly suitable for real-world applications where robustness, accuracy as well as low inference time on high-resolution fundus images is required.