基于人体组织密度的智能掩码，提升3D医学图像建模效果
HU-based Foreground Masking for 3D Medical Masked Image Modeling
- 根据亨氏单位（HU）分布选择性掩码器官区域，忽略空气和液体等无诊断价值部分
- 在5个公开数据集上分割性能显著提升，最高Dice达92.43%
- 适合需要高精度医学图像分割的研究者，尤其关注肺、脑、腹部等器官
虽然掩码图像建模（MIM）已革新计算机视觉领域，但在3D医学图像计算中的应用受限于随机掩码策略，该策略忽略了解剖结构的密度特征。为此，本文提出一种基于亨氏单位（HU）的前景掩码方法，利用组织密度分布聚焦于实质性脏器，排除空气、液体等缺乏诊断意义的非组织区域。在五个公开3D医学影像数据集上的实验表明，该方法在分割质量与Dice分数上均有显著提升：BTCV达~84.64%，Flare22达~92.43%，MM-WHS达~90.67%，Amos22达~88.64%，BraTS达~78.55%。结果凸显了面向领域的MIM的重要性，为医学图像分割的表征学习提供了新方向。代码已开源：github.com/AISeedHub/SubFore/
原文摘要 · Abstract (English)
While Masked Image Modeling (MIM) has revolutionized fields of computer vision, its adoption in 3D medical image computing has been limited by the use of random masking, which overlooks the density of anatomical objects. To address this limitation, we enhance the pretext task with a simple yet effective masking strategy. Leveraging Hounsfield Unit (HU) measurements, we implement an HU-based Foreground Masking, which focuses on the intensity distribution of visceral organs and excludes non-tissue regions, such as air and fluid, that lack diagnostically meaningful features. Extensive experiments on five public 3D medical imaging datasets demonstrate that our masking consistently improves performance, both in quality of segmentation and Dice score (BTCV:~84.64\%, Flare22:~92.43\%, MM-WHS:~90.67\%, Amos22:~88.64\%, BraTS:~78.55\%). These results underscore the importance of domain-centric MIM and suggest a promising direction for representation learning in medical image segmentation. Implementation is available at github.com/AISeedHub/SubFore/.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。