Hard Exudate Segmentation Supplemented by Super-Resolution with Multi-scale Attention Fusion Module

TL;DR

This paper introduces SS-MAF, a novel hard exudate segmentation method that leverages super-resolution and a multi-scale attention fusion module to improve tiny lesion detection and boundary accuracy in retinal images.

Contribution

The paper proposes a dual-stream framework with a Multi-scale Attention Fusion module and region mutual information loss, enhancing tiny lesion and boundary detection in hard exudate segmentation.

Findings

Achieves ≥3% higher dice and recall on E-Ophtha dataset.

Performs competitively with low-resolution inputs.

Effectively detects tiny lesions and boundaries.

Abstract

Hard exudates (HE) is the most specific biomarker for retina edema. Precise HE segmentation is vital for disease diagnosis and treatment, but automatic segmentation is challenged by its large variation of characteristics including size, shape and position, which makes it difficult to detect tiny lesions and lesion boundaries. Considering the complementary features between segmentation and super-resolution tasks, this paper proposes a novel hard exudates segmentation method named SS-MAF with an auxiliary super-resolution task, which brings in helpful detailed features for tiny lesion and boundaries detection. Specifically, we propose a fusion module named Multi-scale Attention Fusion (MAF) module for our dual-stream framework to effectively integrate features of the two tasks. MAF first adopts split spatial convolutional (SSC) layer for multi-scale features extraction and then utilize attention mechanism for features fusion of the two tasks. Considering pixel dependency, we introduce region mutual information (RMI) loss to optimize MAF module for tiny lesions and boundary detection. We evaluate our method on two public lesion datasets, IDRiD and E-Ophtha. Our method shows competitive performance with low-resolution inputs, both quantitatively and qualitatively. On E-Ophtha dataset, the method can achieve $\geq3\%$ higher dice and recall compared with the state-of-the-art methods.