arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.04768cs.CV
收藏

基于随机振幅谱合成的单源域泛化医学图像分割

Medical Image Segmentation via Single-Source Domain Generalization with Random Amplitude Spectrum Synthesis

School of Mechanical, Electrical & Information Engineering, Shandong University(山东大学机电与信息工程学院);Shenzhen Research Institute of Shandong University(山东大学深圳研究院);School of Computer Science and Technology, Shandong University of Finance and Economics(山东财经大学计算机科学与技术学院);Shandong Provincial Key Laboratory of Digital Media Technology(山东省数字媒体技术重点实验室)

Qiang Qiao, Wenyu Wang, Meixia Qu, Kun Su, Bin Jiang, Qiang Guo

AI总结:

针对医学图像分割中的单源域泛化问题,提出RASS频率增强及随机掩码洗牌重建框架RAS^4DG,在胎儿脑和眼底图像上取得更优分割性能。

摘要:

医学图像分割领域因临床数据集中的域偏移而面临域泛化(DG)挑战。医学数据稀缺和隐私问题进一步加剧了DG挑战。传统单源域泛化(SSDG)方法主要依赖堆叠数据增强技术来最小化域差异。本文提出随机振幅谱合成(RASS)作为医学图像的训练增强方法。RASS通过从频率视角模拟分布变化来增强模型泛化能力。该策略通过施加依赖于振幅的扰动引入变化性,以确保广泛覆盖潜在的域变化。此外,我们提出随机掩码洗牌和重建组件,可增强骨干网络处理结构信息的能力,并提高对域内和跨域变化的恢复力。所提出的用于单源域泛化的随机振幅谱合成(RAS^4DG)在三维胎儿脑图像和二维眼底摄影上得到验证,与其他SSDG模型相比实现了更优的DG分割性能。

英文摘要:

The field of medical image segmentation is challenged by domain generalization (DG) due to domain shifts in clinical datasets. The DG challenge is exacerbated by the scarcity of medical data and privacy concerns. Traditional single-source domain generalization (SSDG) methods primarily rely on stacking data augmentation techniques to minimize domain discrepancies. In this paper, we propose Random Amplitude Spectrum Synthesis (RASS) as a training augmentation for medical images. RASS enhances model generalization by simulating distribution changes from a frequency perspective. This strategy introduces variability by applying amplitude-dependent perturbations to ensure broad coverage of potential domain variations. Furthermore, we propose random mask shuffle and reconstruction components, which can enhance the ability of the backbone to process structural information and increase resilience intra- and cross-domain changes. The proposed Random Amplitude Spectrum Synthesis for Single-Source Domain Generalization (RAS^4DG) is validated on 3D fetal brain images and 2D fundus photography, and achieves an improved DG segmentation performance compared to other SSDG models.

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。

↑