GCLR: A self-supervised representation learning pretext task for glomerular filtration barrier segmentation in TEM images.

Autor: Lin G; School of Biomedical Engineering, Southern Medical University, Guangzhou, 510515, China; Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, 510515, China; Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, 510515, China., Zhang Z; School of Biomedical Engineering, Southern Medical University, Guangzhou, 510515, China; Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, 510515, China; Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, 510515, China., Long K; School of Biomedical Engineering, Southern Medical University, Guangzhou, 510515, China; Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, 510515, China; Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, 510515, China., Zhang Y; School of Biomedical Engineering, Southern Medical University, Guangzhou, 510515, China; Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, 510515, China; Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, 510515, China., Lu Y; Central Laboratory, Southern Medical University, Guangzhou, 510515, China., Geng J; Department of Pathology, School of Basic Medical Sciences, Southern Medical University, Guangzhou, 510515, China; Guangzhou Huayin Medical Laboratory Center, Guangzhou, 510515, China., Zhou Z; Central Laboratory, Southern Medical University, Guangzhou, 510515, China., Feng Q; School of Biomedical Engineering, Southern Medical University, Guangzhou, 510515, China; Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, 510515, China; Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, 510515, China., Lu L; School of Biomedical Engineering, Southern Medical University, Guangzhou, 510515, China; Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, 510515, China; Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, 510515, China. Electronic address: ljlubme@gmail.com., Cao L; School of Biomedical Engineering, Southern Medical University, Guangzhou, 510515, China; Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, 510515, China; Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology, Southern Medical University, Guangzhou, 510515, China. Electronic address: caolei@smu.edu.cn.
Jazyk: angličtina
Zdroj: Artificial intelligence in medicine [Artif Intell Med] 2023 Dec; Vol. 146, pp. 102720. Date of Electronic Publication: 2023 Nov 17.
DOI: 10.1016/j.artmed.2023.102720
Abstrakt: Automatic segmentation of the three substructures of glomerular filtration barrier (GFB) in transmission electron microscopy (TEM) images holds immense potential for aiding pathologists in renal disease diagnosis. However, the labor-intensive nature of manual annotations limits the training data for a fully-supervised deep learning model. Addressing this, our study harnesses self-supervised representation learning (SSRL) to utilize vast unlabeled data and mitigate annotation scarcity. Our innovation, GCLR, is a hybrid pixel-level pretext task tailored for GFB segmentation, integrating two subtasks: global clustering (GC) and local restoration (LR). GC captures the overall GFB by learning global context representations, while LR refines three substructures by learning local detail representations. Experiments on 18,928 unlabeled glomerular TEM images for self-supervised pre-training and 311 labeled ones for fine-tuning demonstrate that our proposed GCLR obtains the state-of-the-art segmentation results for all three substructures of GFB with the Dice similarity coefficient of 86.56 ± 0.16%, 75.56 ± 0.36%, and 79.41 ± 0.16%, respectively, compared with other representative self-supervised pretext tasks. Our proposed GCLR also outperforms the fully-supervised pre-training methods based on the three large-scale public datasets - MitoEM, COCO, and ImageNet - with less training data and time.
Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
(Copyright © 2023. Published by Elsevier B.V.)
Databáze: MEDLINE
Externí odkaz: