Browse by author
Lookup NU author(s): Dr Duo Li
This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).
Learning from AI-generated annotations is wellrecognized as a key advance of deep learning techniques in medical image segmentation. Towards this direction, in this paper, we investigate two questions: (1) how to accurately measure loss value on AI-generated annotations that often contain errors and (2) how to effectively update model's parameters when the loss value is no longer a correct supervision for medical image segmentation. The main results are that (1) 'error-tolerant' loss functions exist and (2) 'cross-training', updating the model using data with a small loss of its 'twin' model, can tolerate the loss function to some extent. Per the main results, we yet derived a robust training algorithm, called confidence regularized coteaching, that helps deep models to combat annotation errors in medical image segmentation. This algorithm simultaneously trains two 'twin' segmentation models and updates model's parameters by cross-training with disagreement confident data that are predicted differently by the two models, thereby being able to learning from data with annotation errors. The empirical evidence from a publicly available dataset shows that this new algorithm works better on combating annotation errors than existing methods for medical image segmentation, opening the opportunity to use AI-generated annotations to train segmentation model for medical image segmentation.
Author(s): Song Y, Liu Y, Lin Z, Zhou J, Li D, Zhou T, Leung M-F
Publication type: Article
Publication status: Published
Journal: IEEE Transactions on Consumer Electronics
Year: 2025
Volume: 71
Issue: 1
Pages: 1473-1481
Print publication date: 01/02/2025
Online publication date: 04/10/2024
Acceptance date: 27/09/2024
Date deposited: 24/10/2024
ISSN (print): 0098-3063
ISSN (electronic): 1558-4127
Publisher: IEEE
URL: https://doi.org/10.1109/TCE.2024.3474037
DOI: 10.1109/TCE.2024.3474037
ePrints DOI: 10.57711/7zkd-sj25
Altmetrics provided by Altmetric
