DMSPS: Dynamically mixed soft pseudo-label supervision for scribble-supervised medical image segmentation

被引:3
|
作者
Han, Meng [1 ]
Luo, Xiangde [1 ,3 ]
Xie, Xiangjiang [1 ]
Liao, Wenjun [2 ,4 ]
Zhang, Shichuan [2 ]
Song, Tao [5 ]
Wang, Guotai [1 ,3 ]
Zhang, Shaoting [1 ,3 ]
机构
[1] Univ Elect Sci & Technol China, Sch Mech & Elect Engn, Chengdu, Peoples R China
[2] Sichuan Canc Hosp & Inst, Sichuan Canc Ctr, Dept Radiat Oncol, Chengdu, Peoples R China
[3] Shanghai Artificial Intelligence Lab, Shanghai, Peoples R China
[4] Univ Elect Sci & Technol China, Sch Med, Chengdu, Peoples R China
[5] SenseTime Res, Shanghai, Peoples R China
关键词
Weakly-supervised learning; Scribble annotation; Soft pseudo-label; Uncertainty; Annotation expansion;
D O I
10.1016/j.media.2024.103274
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
High performance of deep learning on medical image segmentation rely on large-scale pixel-level dense annotations, which poses a substantial burden on medical experts due to the laborious and time-consuming annotation process, particularly for 3D images. To reduce the labeling cost as well as maintain relatively satisfactory segmentation performance, weakly-supervised learning with sparse labels has attained increasing attentions. In this work, we present a scribble-based framework for medical image segmentation, called Dynamically Mixed Soft Pseudo-label Supervision (DMSPS). Concretely, we extend a backbone with an auxiliary decoder to form a dual-branch network to enhance the feature capture capability of the shared encoder. Considering that most pixels do not have labels and hard pseudo-labels tend to be over-confident to result in poor segmentation, we propose to use soft pseudo-labels generated by dynamically mixing the decoders' predictions as auxiliary supervision. To further enhance the model's performance, we adopt a two-stage approach where the sparse scribbles are expanded based on predictions with low uncertainties from the first- stage model, leading to more annotated pixels to train the second-stage model. Experiments on ACDC dataset for cardiac structure segmentation, WORD dataset for 3D abdominal organ segmentation and BraTS2020 dataset for 3D brain tumor segmentation showed that: (1) compared with the baseline, our method improved the average DSC from 50.46% to 89.51%, from 75.46% to 87.56% and from 52.61% to 76.53% on the three datasets, respectively; (2) DMSPS achieved better performance than five state-of-the-art scribble-supervised segmentation methods, and is generalizable to different segmentation backbones. The code is available online at: https://github.com/HiLab-git/DMSPS.
引用
收藏
页数:14
相关论文
共 50 条
  • [1] Scribble-Supervised Medical Image Segmentation via Dual-Branch Network and Dynamically Mixed Pseudo Labels Supervision
    Luo, Xiangde
    Hu, Minhao
    Liao, Wenjun
    Zhai, Shuwei
    Song, Tao
    Wang, Guotai
    Zhang, Shaoting
    MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION, MICCAI 2022, PT I, 2022, 13431 : 528 - 538
  • [2] Scribble-supervised medical image segmentation based on dynamically generated pseudo labels via multi-scale superpixels
    Li, Zhixun
    Fang, Jiancheng
    Qiu, Ruiyun
    Gong, Huiling
    BIOMEDICAL SIGNAL PROCESSING AND CONTROL, 2025, 105
  • [3] Semi-supervised medical imaging segmentation with soft pseudo-label fusion
    Xiaoqiang Li
    Yuanchen Wu
    Songmin Dai
    Applied Intelligence, 2023, 53 : 20753 - 20765
  • [4] Semi-supervised medical imaging segmentation with soft pseudo-label fusion
    Li, Xiaoqiang
    Wu, Yuanchen
    Dai, Songmin
    APPLIED INTELLIGENCE, 2023, 53 (18) : 20753 - 20765
  • [5] Non-iterative scribble-supervised learning with pacing pseudo-masks for medical image segmentation
    Yang, Zefan
    Lin, Di
    Ni, Dong
    Wang, Yi
    EXPERT SYSTEMS WITH APPLICATIONS, 2024, 238
  • [6] ScribbleVC: Scribble-supervised Medical Image Segmentation with Vision-Class Embedding
    Li, Zihan
    Zheng, Yuan
    Luo, Xiangde
    Shan, Dandan
    Hong, Qingqi
    PROCEEDINGS OF THE 31ST ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA, MM 2023, 2023, : 3384 - 3393
  • [7] BLPSeg: Balance the Label Preference in Scribble-Supervised Semantic Segmentation
    Wang, Yude
    Zhang, Jie
    Kan, Meina
    Shan, Shiguang
    Chen, Xilin
    IEEE TRANSACTIONS ON IMAGE PROCESSING, 2023, 32 : 4921 - 4934
  • [8] Pseudo-label Guided Contrastive Learning for Semi-supervised Medical Image Segmentation
    Basak, Hritam
    Yin, Zhaozheng
    2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2023, : 19786 - 19797
  • [9] Non-iterative scribble-supervised learning with pacing pseudo-masks for medical image segmentation[Formula presented]
    Yang, Zefan
    Lin, Di
    Ni, Dong
    Wang, Yi
    Expert Systems with Applications, 2024, 238
  • [10] Uncertainty-aware pseudo-label and consistency for semi-supervised medical image segmentation
    Lu, Liyun
    Yin, Mengxiao
    Fu, Liyao
    Yang, Feng
    BIOMEDICAL SIGNAL PROCESSING AND CONTROL, 2023, 79