Joint Medical Image Enhancement and Segmentation with Diffusion-based Symbiotic Information Interaction

2026-06-30

AI Summary

This work addresses the challenge of accurate lesion and anatomical structure segmentation in low-quality medical images—such as MRI, CT, and ultrasound—where limited resolution severely hinders performance. Existing approaches typically treat image enhancement and segmentation as separate tasks, preventing synergistic optimization. To overcome this limitation, the authors propose DiSIINet, a dual-branch framework built upon the denoising diffusion implicit model (DDIM) that jointly performs enhancement and segmentation. The key innovation lies in the newly introduced Symbiotic Information Interaction (SII) mechanism, which enables dynamic, feature-level fusion between the two tasks via cross-attention during the reverse diffusion process. By departing from conventional serial or isolated processing paradigms, DiSIINet achieves significant improvements in both image enhancement quality and segmentation accuracy across multimodal medical imaging datasets, while maintaining efficient inference and stable outputs.

Abstract

Image quality is critical for accurate medical diagnosis. However, MRI, CT, and ultrasound images are often of low resolution and quality due to cost constraints, complicating the visualization of key anatomical structures and lesions. While such limitations are common in practice, traditional methods treat image enhancement as a separate preprocessing step, failing to fully leverage its potential synergy with image segmentation. To address this, we propose DiSIINet (Diffusion-based Symbiotic Information Interaction Network), which is built on the principle that enhancement and segmentation should mutually reinforce each other in a unified model. Based on Denoising Diffusion Implicit Models (DDIM), DiSIINet integrates an enhancement branch and a segmentation branch. These branches interact through a novel Symbiotic Information Interaction (SII) module, which facilitates dynamic, feature-level information exchange via cross-attention during the reverse diffusion process. This design enables both tasks to iteratively improve each other. The DDIM backbone ensures high-quality output and efficient inference through deterministic sampling. Experiments on multi-modal medical datasets (MRI, CT, ultrasound) show that DiSIINet achieves significant performance improvements compared to sequential or independent enhancement and segmentation approaches. The code is available at: https://github.com/Reconsider80/DiSIINet.

Problem

• medical image enhancement
• image segmentation
• low-quality medical images
• synergistic interaction
• multi-modal imaging

Innovation

• Diffusion Model
• Symbiotic Information Interaction
• Joint Enhancement and Segmentation
• Cross-Attention
• Medical Image Analysis