Search papers, labs, and topics across Lattice.
The paper introduces FOSCU, a 3D latent diffusion model with ControlNet, to generate synthetic MRI volumes and segmentation labels conditioned on segmentation maps. This approach aims to alleviate data scarcity and annotation costs in medical image segmentation. Experiments on abdominal MRI scans demonstrate that training 3D U-Nets with combined real and synthetic data improves the mean Dice score by 0.67% and reduces FID by 36.4% compared to training with real data alone.
Synthetic MRI data, generated by a segmentation-conditioned diffusion model, can measurably improve the performance of 3D U-Nets for hepatic segmentation.
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 Fr茅chet Inception Distance (FID), reflecting enhanced image fidelity.
NVIDIA