arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.23845cs.CV
收藏

3D-LLDM:基于标签的3D潜在扩散模型用于提高肝结构分割的高分辨率合成MR影像

3D-LLDM: Label-Guided 3D Latent Diffusion Model for Improving High-Resolution Synthetic MR Imaging in Hepatic Structure Segmentation

OUTTA, Republic of Korea(韩国OUTTA);Stanford University(斯坦福大学);Seoul National University(首尔国立大学);Sangmyung University(Sangmyung大学);Chung-Ang University(Chung-Ang大学);Samsung Medical Center(三星医疗中心);NVIDIA;Sungkyunkwan University School of Medicine(成均馆大学医学院);Samsung Changwon Hospital(三星昌原医院)
Kyeonghun Kim, Jaehyeok Bae, Youngung Han, Joo Young Bae, Seoyoung Ju, Junsu Lim, Gyeongmin Kim, Nam-Joon Kim, Woo Kyoung Jeong, Ken Ying-Kai Liao, Won Jae Lee,… 展开作者

Kyeonghun Kim, Jaehyeok Bae, Youngung Han, Joo Young Bae, Seoyoung Ju, Junsu Lim, Gyeongmin Kim, Nam-Joon Kim, Woo Kyoung Jeong, Ken Ying-Kai Liao, Won Jae Lee, Pa Hong, Hyuk-Jae Lee

AI总结:

本文提出3D-LLDM模型,通过标签引导生成高质量合成MR影像,提升肝结构分割的精度,实验表明其在数据增强中提升了肝细胞癌分割的Dice得分。

摘要:

深度学习和生成模型正迅速发展,合成数据越来越多地被集成到训练管道中用于下游分析任务。然而,在医学影像中,其应用受限于可靠标注数据集的稀缺性。为了解决这一限制,我们提出了3D-LLDM,一种基于标签的3D潜在扩散模型,能够生成高质量的合成磁共振(MR)体积及其对应的解剖分割掩码。我们的方法利用增强的Gd-EOB-DTPA对比剂的肝胆期MR图像,推导出肝脏、门静脉、肝静脉和肝细胞癌的结构掩码,然后通过基于ControlNet的架构进行体积分合成。在720个真实的临床肝胆期MR扫描上训练,3D-LLDM实现了Fréchet Inception Distance(FID)为28.31,比GANs提升了70.9%,比最先进的扩散基线提高了26.7%。当用于数据增强时,合成体积在五个CNN架构上将肝细胞癌分割的Dice得分提升高达11.153%。

英文摘要:

Deep learning and generative models are advancing rapidly, with synthetic data increasingly being integrated into training pipelines for downstream analysis tasks. However, in medical imaging, their adoption remains constrained by the scarcity of reliable annotated datasets. To address this limitation, we propose 3D-LLDM, a label-guided 3D latent diffusion model that generates high-quality synthetic magnetic resonance (MR) volumes with corresponding anatomical segmentation masks. Our approach uses hepatobiliary phase MR images enhanced with the Gd-EOB-DTPA contrast agent to derive structural masks for the liver, portal vein, hepatic vein, and hepatocellular carcinoma, which then guide volumetric synthesis through a ControlNet-based architecture. Trained on 720 real clinical hepatobiliary phase MR scans from Samsung Medical Center, 3D-LLDM achieves a Fréchet Inception Distance (FID) of 28.31, improving over GANs by 70.9% and over state-of-the-art diffusion baselines by 26.7%. When used for data augmentation, the synthetic volumes improve hepatocellular carcinoma segmentation by up to 11.153% Dice score across five CNN architectures.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑