Meta semi-supervised medical image segmentation with label hierarchy

被引:3
作者
Xu, Hai [1 ]
Xie, Hongtao [1 ]
Tan, Qingfeng [2 ]
Zhang, Yongdong [1 ]
机构
[1] Univ Sci & Technol China, Sch Informat Sci & Technol, Hefei 230026, Anhui, Peoples R China
[2] Guangzhou Univ, Cyberspace Inst Adv Technol, Guangzhou 511442, Guangdong, Peoples R China
关键词
Medical image segmentation; Semi-supervised learning; Consistency regularization; Domain generalization;
D O I
10.1007/s13755-023-00222-1
中图分类号
R-058 [];
学科分类号
摘要
Semi-supervised learning (SSL) has attracted increasing attention in medical image segmentation, where the mainstream usually explores perturbation-based consistency as a regularization to leverage unlabelled data. However, unlike directly optimizing segmentation task objectives, consistency regularization is a compromise by incorporating invariance towards perturbations, and inevitably suffers from noise in self-predicted targets. The above issues result in a knowledge gap between supervised guidance and unsupervised regularization. To bridge the knowledge gap, this work proposes a meta-based semi-supervised segmentation framework with the exploitation of label hierarchy. Two main prominent components named Divide and Generalize, and Label Hierarchy, are built in this work. Concretely, rather than merging all knowledge indiscriminately, we dynamically divide consistency regularization from supervised guidance as different domains. Then, a domain generalization technique is introduced with a meta-based optimization objective which ensures the update on supervised guidance should generalize to the consistency regularization, thereby bridging the knowledge gap. Furthermore, to alleviate the negative impact of noise in self-predicted targets, we propose to distill the noisy pixel-level consistency by exploiting label hierarchy and extracting hierarchical consistencies. Comprehensive experiments on two public medical segmentation benchmarks demonstrate the superiority of our framework to other semi-supervised segmentation methods, with new state-of-the-art results.
引用
收藏
页数:15
相关论文
共 50 条
  • [31] Multidimensional perturbed consistency learning for semi-supervised medical image segmentation
    Yuan, Enze
    Zhao, Bin
    Qin, Xiao
    Ding, Shuxue
    [J]. INTERNATIONAL JOURNAL OF IMAGING SYSTEMS AND TECHNOLOGY, 2024, 34 (03)
  • [32] Semi-supervised medical image segmentation network based on mutual learning
    Sun, Junmei
    Wang, Tianyang
    Wang, Meixi
    Li, Xiumei
    Xu, Yingying
    [J]. MEDICAL PHYSICS, 2025, 52 (03) : 1589 - 1600
  • [33] Cross Task Temporal Consistency for Semi-supervised Medical Image Segmentation
    Jeevan, Govind
    Pawan, S. J.
    Rajan, Jeny
    [J]. MACHINE LEARNING IN MEDICAL IMAGING, MLMI 2022, 2022, 13583 : 140 - 150
  • [34] A Lightweight Deep Semi-supervised Student Model for Medical Image Segmentation
    Le Dinh Huynh
    Truong Cong Doan
    Phan Duy Hung
    [J]. COOPERATIVE DESIGN, VISUALIZATION, AND ENGINEERING, CDVE 2024, 2024, 15158 : 233 - 242
  • [35] Semi-Supervised Medical Image Segmentation with Co-Distribution Alignment
    Wang, Tao
    Huang, Zhongzheng
    Wu, Jiawei
    Cai, Yuanzheng
    Li, Zuoyong
    [J]. BIOENGINEERING-BASEL, 2023, 10 (07):
  • [36] A Novel Perturbation Consistency Framework in Semi-Supervised Medical Image Segmentation
    Ma, Xiaoxuan
    Lian, Kuncheng
    Sui, Dong
    [J]. APPLIED SCIENCES-BASEL, 2024, 14 (18):
  • [37] Enhancing semi-supervised medical image segmentation with boundary awareness and Contrastive
    Lin, Zhikang
    Li, Zhenteng
    Su, Jiawei
    Li, Lei
    Lian, Sheng
    [J]. BIOMEDICAL SIGNAL PROCESSING AND CONTROL, 2025, 103
  • [38] Dual-decoder data decoupling training for semi-supervised medical image segmentation
    Wang, Bing
    Huang, Taifeng
    Yang, Shuo
    Yang, Ying
    Zhai, Junhai
    Zhang, Xin
    [J]. BIOMEDICAL SIGNAL PROCESSING AND CONTROL, 2025, 100
  • [39] Federated Semi-Supervised Learning for Medical Image Segmentation via Pseudo-Label Denoising
    Qiu, Liang
    Cheng, Jierong
    Gao, Huxin
    Xiong, Wei
    Ren, Hongliang
    [J]. IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, 2023, 27 (10) : 4672 - 4683
  • [40] Own-background contrastive learning guided by pseudo-label for semi-supervised medical image segmentation
    Fan, Huijie
    Cao, Jinghan
    Chen, Xi'ai
    Lin, Sen
    Polat, Kemal
    Zhou, Jingchun
    [J]. APPLIED SOFT COMPUTING, 2025, 171