Chinese Medical Sciences Journal ›› 2022, Vol. 37 ›› Issue (3): 171-180.doi: 10.24920/004135

• 科学数据共享与重用: 论著 •    下一篇

半监督内镜图像长尾分类

操润楠1,2,方梦捷1,2,李海林3,田捷2,3,4,董迪1,2,*()   

  1. 1中国科学院大学 人工智能学院,北京 100049,中国
    2中国科学院自动化研究所,中国科学院分子影像重点实验室,分子影像北京市重点实验室,复杂系统管理与控制国家重点实验室,北京100190,中国
    3北京航空航天大学 医学科学与工程学院,大数据精准医疗高精尖创新中心,北京100191,中国
    4西安电子科技大学 生命科学技术学院,分子与神经影像教育部工程研究中心,西安710126,中国
  • 收稿日期:2022-06-30 接受日期:2022-09-09 出版日期:2022-09-30 发布日期:2022-09-27
  • 通讯作者: 董迪 E-mail:di.dong@ia.ac.cn

Semi-supervised Long-tail Endoscopic Image Classification

Runnan Cao1,2,Mengjie Fang1,2,Hailing Li3,Jie Tian2,3,4,Di Dong1,2,*()   

  1. 1School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing 100049, China
    2CAS Key Laboratory of Molecular Imaging, Beijing Key Laboratory of Molecular Imaging, the State Key Laboratory of Management and Control for Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
    3Beijing Advanced Innovation Center for Big Data-Based Precision Medicine, School of Engineering Medicine, Beihang University, Beijing 100191, China
    4Engineering Research Center of Molecular and Neuro Imaging of Ministry of Education, School of Life Science and Technology, Xidian University, Xi’an 710126, China
  • Received:2022-06-30 Accepted:2022-09-09 Published:2022-09-30 Online:2022-09-27
  • Contact: Di Dong E-mail:di.dong@ia.ac.cn

摘要:

目的 探索半监督学习算法在内镜图像长尾分类中的应用。
方法 我们在HyperKvasir数据集上探索了半监督的内镜图像长尾分类,该数据集是最大的胃肠道公共数据集,有23个不同的类别。使用基于一致性正则化和伪标签的半监督学习算法FixMatch,在将训练数据集和测试数据集按4:1的比例进行划分后,按照20%、50%和100%的比例抽取有标签的训练样本,以测试在有标签数据有限下的分类性能。
结果 通过微观平均、宏观平均评价指标和马修斯相关系数(Mathews correlation coefficient,MCC)作为总体评价指标来评估分类性能。半监督学习算法在有标签训练数据比例为20%、50%和100%的情况下,MCC分别从0.8761提高到0.8850、0.8983提高到0.8994、0.9075提高到0.9095。在有标签训练数据比例为20%的情况下,半监督学习算法可以提高微观平均和宏观平均的分类性能。对于50%和100%的情况,半监督学习算法可以提高微观平均下的分类性能,但会损害宏观平均的分类性能。通过分析每个类的混淆矩阵和标注偏差,我们发现基于伪标签的半监督学习算法加剧了分类器对头类的偏好,导致头类的性能提高而尾类的性能下降。
结论 半监督学习算法可以提高内镜图像长尾分类的性能,特别是在标签极其有限的情况下,这可能有利于为小医院建立辅助诊断系统。然而,伪标签策略可能会放大类不平衡的影响,从而损害尾部类的分类性能。

关键词: 内镜图像, 人工智能, 半监督学习, 长尾分类, 图像分类

Abstract:

Objective To explore the semi-supervised learning (SSL) algorithm for long-tail endoscopic image classification with limited annotations.
Method We explored semi-supervised long-tail endoscopic image classification in HyperKvasir, the largest gastrointestinal public dataset with 23 diverse classes. Semi-supervised learning algorithm FixMatch was applied based on consistency regularization and pseudo-labeling. After splitting the training dataset and the test dataset at a ratio of 4:1, we sampled 20%, 50%, and 100% labeled training data to test the classification with limited annotations.
Results The classification performance was evaluated by micro-average and macro-average evaluation metrics, with the Mathews correlation coefficient (MCC) as the overall evaluation. SSL algorithm improved the classification performance, with MCC increasing from 0.8761 to 0.8850, from 0.8983 to 0.8994, and from 0.9075 to 0.9095 with 20%, 50%, and 100% ratio of labeled training data, respectively. With a 20% ratio of labeled training data, SSL improved both the micro-average and macro-average classification performance; while for the ratio of 50% and 100%, SSL improved the micro-average performance but hurt macro-average performance. Through analyzing the confusion matrix and labeling bias in each class, we found that the pseudo-based SSL algorithm exacerbated the classifier’s preference for the head class, resulting in improved performance in the head class and degenerated performance in the tail class.
Conclusion SSL can improve the classification performance for semi-supervised long-tail endoscopic image classification, especially when the labeled data is extremely limited, which may benefit the building of assisted diagnosis systems for low-volume hospitals. However, the pseudo-labeling strategy may amplify the effect of class imbalance, which hurts the classification performance for the tail class.

Key words: endoscopic image, artificial intelligence, semi-supervised learning, long-tail distribution, image classification

基金资助: 中国科学院战略重点研究计划(XDB 38040200);中国国家自然科学基金(82022036);中国国家自然科学基金(91959130);中国国家自然科学基金(81971776);中国国家自然科学基金(62027901);中国国家自然科学基金(81930053);北京市自然科学基金(Z20J00105);国家重点研发计划(2017YFA0205200);中国科学院青年创新促进会(Y2021049)

Copyright © 2018 Chinese Academy of Medical Sciences. All right reserved.
 
www.cmsj.cams.cn
京公安备110402430088 京ICP备06002729号-1  Powered by Magtech.

Supervised by National Health Commission of the People's Republic of China

9 Dongdan Santiao, Dongcheng district, Beijing, 100730 China

Tel: 86-10-65105897  Fax:86-10-65133074 

E-mail: cmsj@cams.cn  www.cmsj.cams.cn

Copyright © 2018 Chinese Academy of Medical Sciences

All right reserved.

京公安备110402430088  京ICP备06002729号-1