HRDecoder: High-Resolution Decoder Network for Fundus Image Lesion Segmentation

TL;DR

HRDecoder is a novel high-resolution decoder network that enhances fundus lesion segmentation accuracy by efficiently capturing local details and multi-scale features, balancing performance with computational cost.

Contribution

It introduces a simple high-resolution decoder with fusion modules that improves segmentation accuracy while maintaining reasonable memory and speed.

Findings

Improved segmentation accuracy on IDRiD and DDR datasets.

Effective balance between high-resolution detail capture and computational efficiency.

Code availability facilitates reproducibility and further research.

Abstract

High resolution is crucial for precise segmentation in fundus images, yet handling high-resolution inputs incurs considerable GPU memory costs, with diminishing performance gains as overhead increases. To address this issue while tackling the challenge of segmenting tiny objects, recent studies have explored local-global fusion methods. These methods preserve fine details using local regions and capture long-range context information from downscaled global images. However, the necessity of multiple forward passes inevitably incurs significant computational overhead, adversely affecting inference speed. In this paper, we propose HRDecoder, a simple High-Resolution Decoder network for fundus lesion segmentation. It integrates a high-resolution representation learning module to capture fine-grained local features and a high-resolution fusion module to fuse multi-scale predictions. Our method effectively improves the overall segmentation accuracy of fundus lesions while consuming reasonable memory and computational overhead, and maintaining satisfying inference speed. Experimental results on the IDRiD and DDR datasets demonstrate the effectiveness of our method. Code is available at https://github.com/CVIU-CSU/HRDecoder.