SGW-GAN: Sliced Gromov-Wasserstein Guided GANs for Retinal Fundus Image Enhancement

2026-01-19

AI Summary

Fundus images are frequently degraded by noise, artifacts, and uneven illumination, and existing enhancement methods often distort the intra-class geometric structures of pathological regions, thereby compromising downstream diagnostic performance. To address this issue, this work proposes the SGW-GAN framework, which introduces the Sliced Gromov–Wasserstein (SGW) distance—used here for the first time in unpaired fundus image enhancement—to efficiently approximate the Gromov–Wasserstein distance via random projections. This approach enables distribution alignment while preserving intra-class structural fidelity. Experiments on public datasets demonstrate that the proposed method significantly improves visual quality, outperforms state-of-the-art techniques in diabetic retinopathy grading accuracy, and achieves the lowest GW discrepancy across all disease categories, confirming its clinical fidelity and computational efficiency.

Technology Category

• Computer Vision: Other Foundations of Computer Vision
• Machine Learning: Graph-based Machine Learning
• Search and Optimization: Distributed Search

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Retinal fundus photography is indispensable for ophthalmic screening and diagnosis, yet image quality is often degraded by noise, artifacts, and uneven illumination. Recent GAN- and diffusion-based enhancement methods improve perceptual quality by aligning degraded images with high-quality distributions, but our analysis shows that this focus can distort intra-class geometry: clinically related samples become dispersed, disease-class boundaries blur, and downstream tasks such as grading or lesion detection are harmed. The Gromov Wasserstein (GW) discrepancy offers a principled solution by aligning distributions through internal pairwise distances, naturally preserving intra-class structure, but its high computational cost restricts practical use. To overcome this, we propose SGW-GAN, the first framework to incorporate Sliced GW (SGW) into retinal image enhancement. SGW approximates GW via random projections, retaining relational fidelity while greatly reducing cost. Experiments on public datasets show that SGW-GAN produces visually compelling enhancements, achieves superior diabetic retinopathy grading, and reports the lowest GW discrepancy across disease labels, demonstrating both efficiency and clinical fidelity for unpaired medical image enhancement.

Problem

• Retinal Fundus Image Enhancement
• Gromov-Wasserstein Discrepancy
• Intra-class Geometry Preservation
• Unpaired Medical Image Enhancement
• Downstream Task Degradation

Innovation

• Sliced Gromov-Wasserstein
• Retinal Fundus Image Enhancement
• Unpaired Image-to-Image Translation
• Intra-class Geometry Preservation
• Medical Image GAN