Super-Resolution Enhancement of Medical Images Based on Diffusion Model: An Optimization Scheme for Low-Resolution Gastric Images

TL;DR

Capsule endoscopy yields low-resolution GI imagery that hampers diagnosis. The authors adopt a diffusion-based SR framework (SR3) trained on the HyperKvasir dataset to upsample LR inputs to HR outputs while preserving anatomical fidelity and reducing hallucinations common to GANs. They present two generations of SR3 models with architectural and optimization enhancements, achieving PSNR up to 29.3 dB and SSIM up to 0.71, and propose latent-diffusion via VAE as a future direction for faster, semantically coherent inference. Qualitative results show preserved mucosal boundaries and vascular patterns, supporting diffusion-based SR as a promising path for non-invasive medical image enhancement. The work lays groundwork for real-time deployment and broader clinical validation in capsule endoscopy workflows.

Abstract

Capsule endoscopy has enabled minimally invasive gastrointestinal imaging, but its clinical utility is limited by the inherently low resolution of captured images due to hardware, power, and transmission constraints. This limitation hampers the identification of fine-grained mucosal textures and subtle pathological features essential for early diagnosis. This work investigates a diffusion-based super-resolution framework to enhance capsule endoscopy images in a data-driven and anatomically consistent manner. We adopt the SR3 (Super-Resolution via Repeated Refinement) framework built upon Denoising Diffusion Probabilistic Models (DDPMs) to learn a probabilistic mapping from low-resolution to high-resolution images. Unlike GAN-based approaches that often suffer from training instability and hallucination artifacts, diffusion models provide stable likelihood-based training and improved structural fidelity. The HyperKvasir dataset, a large-scale publicly available gastrointestinal endoscopy dataset, is used for training and evaluation. Quantitative results demonstrate that the proposed method significantly outperforms bicubic interpolation and GAN-based super-resolution methods such as ESRGAN, achieving PSNR of 27.5 dB and SSIM of 0.65 for a baseline model, and improving to 29.3 dB and 0.71 with architectural enhancements including attention mechanisms. Qualitative results show improved preservation of anatomical boundaries, vascular patterns, and lesion structures. These findings indicate that diffusion-based super-resolution is a promising approach for enhancing non-invasive medical imaging, particularly in capsule endoscopy where image resolution is fundamentally constrained.

Figures (16)

• Figure 1: Example of low-resolution capsule endoscopy image showing limited detail for diagnostic purposes.
• Figure 2: Visual results from Real-ESRGAN: comparing bicubic input (left) and ESRGAN-enhanced output (right).
• Figure 3: Simplified architecture of the ESRGAN super-resolution network.
• Figure 4: SR3 Super-Resolution Process: Starting with low-resolution input, the model refines the image over multiple iterations using learned noise predictions to enhance resolution.
• Figure 5: SR3 Reverse Steps: Starting with a noisy upsampled input, the model iteratively predicts and removes noise over $T$ timesteps to refine the image and generate a high-resolution output.