Tongue Color Classification in TCM with Noisy Labels via Confident-Learning-Assisted Knowledge Distillation

被引:5
作者
Li, Yanping [1 ,2 ]
Zhuo, Li [1 ,2 ]
Sun, Liangliang [1 ,2 ]
Zhang, Hui [1 ,2 ]
Li, Xiaoguang [1 ,2 ]
Yang, Yang [3 ]
Wei, Wei [3 ,4 ]
机构
[1] Beijing Univ Technol, Beijing Key Lab Computat Intelligence & Intelligen, Beijing 100124, Peoples R China
[2] Beijing Univ Technol, Fac Informat Technol, Beijing 100124, Peoples R China
[3] China Acad Chinese Med Sci, Wangjing Hosp, Dept Gastroenterol, Beijing 100102, Peoples R China
[4] Beijing Key Lab Tradit Chinese Med Treatment Funct, Beijing 100102, Peoples R China
基金
中国国家自然科学基金;
关键词
Training; Knowledge engineering; Deep learning; Tongue; Image color analysis; Neural networks; Robustness; Traditional Chinese medicine; Tongue color classification; Confident learning; Learning from noisy labels; Knowledge distillation; Channel attention mechanism; ResNet18;
D O I
10.23919/cje.2022.00.040
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Tongue color is an important tongue diagnostic index for traditional Chinese medicine (TCM). Due to the individual experience of TCM experts as well as ambiguous boundaries among the tongue color categories, there often exist noisy labels in annotated samples. Deep neural networks trained with the noisy labeled samples often have poor generalization capability because they easily overfit on noisy labels. A novel framework named confident-learning-assisted knowledge distillation (CLA-KD) is proposed for tongue color classification with noisy labels. In this framework, the teacher network plays two important roles. On the one hand, it performs confident learning to identify, cleanse and correct noisy labels. On the other hand, it learns the knowledge from the clean labels, which will then be transferred to the student network to guide its training. Moreover, we elaborately design a teacher network in an ensemble manner, named E-CA(2)-ResNet18, to solve the unreliability and instability problem resulted from the insufficient data samples. E-CA(2)-ResNet18 adopts ResNet18 as the backbone, and integrates channel attention (CA) mechanism and activate or not activation function together, which facilitates to yield a better performance. The experimental results on three self-established TCM tongue datasets demonstrate that, our proposed CLA-KD can obtain a superior classification accuracy and good robustness with a lower network model complexity, reaching 94.49%, 92.21%, 93.43% on the three tongue image datasets, respectively.
引用
收藏
页码:140 / 150
页数:11
相关论文
共 38 条
  • [1] Algan G, 2020, Arxiv, DOI arXiv:2003.10471
  • [2] Image classification with deep learning in the presence of noisy labels: A survey
    Algan, Gorkem
    Ulusoy, Ilkay
    [J]. KNOWLEDGE-BASED SYSTEMS, 2021, 215
  • [3] [Anonymous], 2018, P C WORKSH NEUR INF
  • [4] [Anonymous], 1988, MACH LEARN, V2, P343
  • [5] Ordinal Regression Methods: Survey and Experimental Study
    Antonio Gutierrez, Pedro
    Perez-Ortiz, Maria
    Sanchez-Monedero, Javier
    Fernandez-Navarro, Francisco
    Hervas-Martinez, Cesar
    [J]. IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2016, 28 (01) : 127 - 146
  • [6] Chen LY, 2016, INT CONF COMP SCI ED, P113, DOI 10.1109/ICCSE.2016.7581565
  • [7] SCA-CNN: Spatial and Channel-wise Attention in Convolutional Networks for Image Captioning
    Chen, Long
    Zhang, Hanwang
    Xiao, Jun
    Nie, Liqiang
    Shao, Jian
    Liu, Wei
    Chua, Tat-Seng
    [J]. 30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017), 2017, : 6298 - 6306
  • [8] On the Efficacy of Knowledge Distillation
    Cho, Jang Hyun
    Hariharan, Bharath
    [J]. 2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, : 4793 - 4801
  • [9] Profiling instances in noise reduction
    Delany, Sarah Jane
    Segata, Nicola
    Mac Namee, Brian
    [J]. KNOWLEDGE-BASED SYSTEMS, 2012, 31 : 28 - 40
  • [10] New label noise injection methods for the evaluation of noise filters
    Garcia, Luis P. F.
    Lehmann, Jens
    de Carvalho, Andre C. P. L. F.
    Lorena, Ana C.
    [J]. KNOWLEDGE-BASED SYSTEMS, 2019, 163 : 693 - 704