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

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Han, Youngung, Kim, Kyeonghun, Ju, Seoyoung, Jean, Yeonju, Cha, Minkyung, Park, Seohyoung, Jung, Hyeonseok, Kim, Nam-Joon, Jeong, Woo Kyoung, Liao, Ken Ying-Kai, Lee, Hyuk-Jae
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915901143515136
author Han, Youngung
Kim, Kyeonghun
Ju, Seoyoung
Jean, Yeonju
Cha, Minkyung
Park, Seohyoung
Jung, Hyeonseok
Kim, Nam-Joon
Jeong, Woo Kyoung
Liao, Ken Ying-Kai
Lee, Hyuk-Jae
author_facet Han, Youngung
Kim, Kyeonghun
Ju, Seoyoung
Jean, Yeonju
Cha, Minkyung
Park, Seohyoung
Jung, Hyeonseok
Kim, Nam-Joon
Jeong, Woo Kyoung
Liao, Ken Ying-Kai
Lee, Hyuk-Jae
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2603_29343
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle FOSCU: Feasibility of Synthetic MRI Generation via Duo-Diffusion Models for Enhancement of 3D U-Nets in Hepatic Segmentation
Han, Youngung
Kim, Kyeonghun
Ju, Seoyoung
Jean, Yeonju
Cha, Minkyung
Park, Seohyoung
Jung, Hyeonseok
Kim, Nam-Joon
Jeong, Woo Kyoung
Liao, Ken Ying-Kai
Lee, Hyuk-Jae
Computer Vision and Pattern Recognition
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.
title FOSCU: Feasibility of Synthetic MRI Generation via Duo-Diffusion Models for Enhancement of 3D U-Nets in Hepatic Segmentation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.29343