J Shanghai Jiaotong Univ Sci ›› 2025, Vol. 30 ›› Issue (4): 800-814.doi: 10.1007/s12204-024-2733-0

• • 上一篇    下一篇

用于半监督医学图像分割的多一致性训练

  

  1. 上海交通大学 仪器科学与工程系,上海200240
  • 收稿日期:2023-06-19 接受日期:2023-10-24 发布日期:2025-07-31

Multi-Consistency Training for Semi-Supervised Medical Image Segmentation

吴昌学,章闻曦,韩佼志,王红雨   

  1. Department of Instrument Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, Chi
  • Received:2023-06-19 Accepted:2023-10-24 Published:2025-07-31

摘要: 医学图像分割是临床应用中的一项重要任务。然而,获得医学图像的标记数据通常具有挑战性。这就提高了半监督学习(SSL)的吸引力。半监督学习是一种只需要少量标记数据的技术。尽管如此,大多数流行的医学图像SSL分割方法要么依赖于单一的一致性训练方法,要么直接微调为自然图像设计的SSL方法。本文提出了一种创新的半监督方法,称为多一致性训练(MCT),用于医学图像分割。我们的方法超越了现有方法的限制,从两个角度考虑一致性:不同上采样方法的输出一致性,以及对中间特征的各种扰动下,同一网络相同数据的输出一致性。为这两种类型的一致性,设计了不同的半监督损失函数。为了增强MCT模型的应用,还开发了一个专用解码器作为神经网络的核心。在息肉数据集和牙科数据集上进行了彻底的实验,并与其他SSL方法进行了严格的比较。实验结果证明了该方法的优越性,实现了更高的分割精度。此外,全面的消融研究和深入的讨论证实了我们的方法在处理复杂的医学图像分割方面的有效性。

关键词: 半监督学习,一致性训练,医学图像分割,中间特征扰动

Abstract: Medical image segmentation is a crucial task in clinical applications. However, obtaining labeled data for medical images is often challenging. This has led to the appeal of semi-supervised learning (SSL), a technique adept at leveraging a modest amount of labeled data. Nonetheless, most prevailing SSL segmentation methods for medical images either rely on the single consistency training method or directly fine-tune SSL methods designed for natural images. In this paper, we propose an innovative semi-supervised method called multi-consistency training (MCT) for medical image segmentation. Our approach transcends the constraints of prior methodologies by considering consistency from a dual perspective: output consistency across different up-sampling methods and output consistency of the same data within the same network under various perturbations to the intermediate features. We design distinct semi-supervised loss regression methods for these two types of consistencies. To enhance the application of our MCT model, we also develop a dedicated decoder as the core of our neural network. Thorough experiments were conducted on the polyp dataset and the dental dataset, rigorously compared against other SSL methods. Experimental results demonstrate the superiority of our approach, achieving higher segmentation accuracy. Moreover, comprehensive ablation studies and insightful discussion substantiate the efficacy of our approach in navigating the intricacies of medical image segmentation.

Key words: semi-supervised learning (SSL), multi-consistency training (MCT), medical image segmentation, intermediate feature perturbation

中图分类号: