A dynamic few-shot learning framework for medical image stream mining based on self-training

EURASIP Journal on Advances in Signal Processing

Research

Wei Zhang¹  Zhengqiang Ye² 

[1] Biomedical Informatics and Statistics Center, School of Public Health, Fudan University, Shanghai, China;Information Center, Eye and ENT Hospital, Fudan University, Shanghai, China;

关键词: Information medicine;    Few-shot learning;    Semantic segmentation;    Data steam mining;   

DOI  :  10.1186/s13634-023-00999-z

 received in 2022-11-22, accepted in 2023-03-01,  发布年份 2023

来源: Springer

【 摘 要 】

Few-shot semantic segmentation (FSS) has been widely used in the field of information medicine and intelligent diagnosis. Due to the high cost of medical data collection and the privacy protection of patients, labeled medical images are difficult to obtain. Compared with other semantic segmentation dataset which can be automatically generated in a large scale, the medical image data tend to be continually generated. Most of the existing FSS techniques require abundant annotated semantic classes for pre-training and cannot deal with its dynamic nature of medical data stream. To deal with this issue, we propose a dynamic few-shot learning framework for medical semantic segmentation, which can fully utilize the features of newly-collected/generated data stream. We introduce a new pseudo-label generation strategy for continuously generating pseudo-labels and avoiding model collapse during self-training. Furthermore, an efficient consistency regularization strategy is proposed to fully utilize the limited data. The proposed framework is iteratively trained on three tasks: abdominal organ segmentation for CT and MRI, and cardiac segmentation for MRI. Experiments results demonstrate significant performance gain on medical data stream mining compared with the baseline method.

【 授权许可】

CC BY   
© The Author(s) 2023

【 图 表 】

Fig. 17

41512_2023_147_Article_IEq107.gif

40517_2023_256_Article_IEq116.gif

Fig. 16

40517_2023_256_Article_IEq115.gif

40517_2023_256_Article_IEq114.gif

Fig. 2

40517_2023_256_Article_IEq113.gif

Fig. 5

Fig. 14

Fig. 1

Fig. 1

Fig. 1

40517_2023_256_Article_IEq110.gif

Fig. 5

40517_2023_256_Article_IEq109.gif

Fig. 1

40517_2023_256_Article_IEq108.gif

Fig. 5

40517_2023_256_Article_IEq107.gif

Fig. 4

Fig. 12

Fig. 3

40517_2023_256_Article_IEq79.gif

40517_2023_256_Article_IEq78.gif

Fig. 1

40517_2023_256_Article_IEq73.gif

Fig. 8

Fig. 5

Fig. 3

Fig. 1

40517_2023_256_Article_IEq64.gif

40517_2023_256_Article_IEq62.gif

Fig. 1

Fig. 3

13731_2023_296_Article_IEq2.gif

40517_2023_256_Article_IEq60.gif

Fig. 1

40517_2023_256_Article_IEq59.gif

Fig. 3

Fig. 4

40517_2023_256_Article_IEq55.gif

Fig. 3

Fig. 3

Fig. 2

40517_2023_256_Article_IEq53.gif

Fig. 2

40517_2023_256_Article_IEq52.gif

40517_2023_256_Article_IEq51.gif

Fig. 2

12888_2023_4880_Article_IEq1.gif

Fig. 1

Fig. 1

40517_2023_256_Article_IEq47.gif

Fig. 4

Fig. 1

40517_2023_256_Article_IEq39.gif

40517_2023_256_Article_IEq38.gif

40517_2023_256_Article_IEq36.gif

12936_2023_4577_Article_IEq66.gif

40517_2023_256_Article_IEq34.gif

Fig. 1

Fig. 2

40517_2023_258_Article_IEq68.gif

40517_2023_258_Article_IEq138.gif

Fig. 5

41116_2023_36_Article_IEq680.gif

41116_2023_36_Article_IEq633.gif

40517_2023_258_Article_IEq8.gif

41116_2023_36_Article_IEq478.gif

41116_2023_36_Article_IEq457.gif