FOSCU: Feasibility of Synthetic MRI Generation via Duo-Diffusion Models for Enhancement of 3D U-Nets in Hepatic Segmentation

2026-03-31

AI Summary

Medical image segmentation is hindered by the scarcity of clinical data and the high cost of annotation, which limits model robustness. To address this challenge, this work proposes the FOSCU framework, which introduces an innovative 3D latent diffusion model—dubbed Duo-Diffusion—augmented with ControlNet to jointly generate high-resolution, anatomically plausible synthetic MRI scans and their corresponding segmentation labels under the guidance of segmentation conditions. This approach ensures spatial consistency and fine-grained anatomical fidelity. The generated synthetic data are integrated into an enhanced 3D U-Net training pipeline to improve segmentation performance. Experiments on a dataset of 720 abdominal MRI scans demonstrate that incorporating synthetic data yields a 0.67% absolute improvement in Dice score and a 36.4% reduction in Fréchet Inception Distance, significantly enhancing both image fidelity and segmentation accuracy.

Abstract

Medical image segmentation faces fundamental challenges including restricted access, costly annotation, and data shortage to clinical datasets through Picture Archiving and Communication Systems (PACS). These systemic barriers significantly impede the development of robust segmentation algorithms. To address these challenges, we propose FOSCU, which integrates Duo-Diffusion, a 3D latent diffusion model with ControlNet that simultaneously generates high-resolution, anatomically realistic synthetic MRI volumes and corresponding segmentation labels, and an enhanced 3D U-Net training pipeline. Duo-Diffusion employs segmentation-conditioned diffusion to ensure spatial consistency and precise anatomical detail in the generated data. Experimental evaluation on 720 abdominal MRI scans shows that models trained with combined real and synthetic data yield a mean Dice score gain of 0.67% over those using only real data, and achieve a 36.4% reduction in Fréchet Inception Distance (FID), reflecting enhanced image fidelity.

Problem

• medical image segmentation
• data scarcity
• costly annotation
• clinical datasets
• PACS

Innovation

• Duo-Diffusion
• synthetic MRI generation
• 3D U-Net
• segmentation-conditioned diffusion
• medical image segmentation