Intelligent Masking: Deep Q-Learning for Context Encoding in Medical Image Analysis

被引:0
|
作者
Bahrami, Mojtaba [1 ,2 ]
Ghorbani, Mahsa [1 ,2 ]
Yeganeh, Yousef [1 ]
Navab, Nassir [1 ,3 ]
机构
[1] Tech Univ Munich, Comp Aided Med Procedures, Munich, Germany
[2] Sharif Univ Technol, Tehran, Iran
[3] Johns Hopkins Univ, Whiting Sch Engn, Baltimore, MD USA
来源
MACHINE LEARNING IN MEDICAL IMAGING, MLMI 2022 | 2022年 / 13583卷
关键词
Self-supervised learning; Context encoding; Reinforcement learning; Medical image analysis;
D O I
10.1007/978-3-031-21014-3_40
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The need for a large amount of labeled data in the supervised setting has led recent studies to utilize self-supervised learning to pre-train deep neural networks using unlabeled data. Many self-supervised training strategies have been investigated especially for medical datasets to leverage the information available in the much fewer unlabeled data. One of the fundamental strategies in image-based self-supervision is context prediction. In this approach, a model is trained to reconstruct the contents of an arbitrary missing region of an image based on its surroundings. However, the existing methods adopt a random and blind masking approach by focusing uniformly on all regions of the images. This approach results in a lot of unnecessary network updates that cause the model to forget the rich extracted features. In this work, we develop a novel self-supervised approach that occludes targeted regions to improve the pre-training procedure. To this end, we propose a reinforcement learning-based agent which learns to intelligently mask input images through deep Q-learning. We show that training the agent against the prediction model can significantly improve the semantic features extracted for downstream classification tasks. We perform our experiments on two public datasets for diagnosing breast cancer in the ultrasound images and detecting lower-grade glioma with MR images. In our experiments, we show that our novel masking strategy advances the learned features according to the performance on the classification task in terms of accuracy, macro F1, and AUROC.
引用
收藏
页码:387 / 396
页数:10
相关论文
共 50 条
  • [21] Deep Spatial Q-Learning for Infectious Disease Control
    Liu, Zhishuai
    Clifton, Jesse
    Laber, Eric B.
    Drake, John
    Fang, Ethan X.
    JOURNAL OF AGRICULTURAL BIOLOGICAL AND ENVIRONMENTAL STATISTICS, 2023, 28 (04) : 749 - 773
  • [22] Deep spectral Q-learning with application to mobile health
    Gao, Yuhe
    Shi, Chengchun
    Song, Rui
    STAT, 2023, 12 (01):
  • [23] Deep Spatial Q-Learning for Infectious Disease Control
    Zhishuai Liu
    Jesse Clifton
    Eric B. Laber
    John Drake
    Ethan X. Fang
    Journal of Agricultural, Biological and Environmental Statistics, 2023, 28 : 749 - 773
  • [24] Autonomous Warehouse Robot using Deep Q-Learning
    Peyas, Ismot Sadik
    Hasan, Zahid
    Tushar, Md Rafat Rahman
    Al Musabbir
    Azni, Raisa Mehjabin
    Siddique, Shahnewaz
    2021 IEEE REGION 10 CONFERENCE (TENCON 2021), 2021, : 857 - 862
  • [25] A Tour of Unsupervised Deep Learning for Medical Image Analysis
    Raza, Khalid
    Singh, Nripendra Kumar
    CURRENT MEDICAL IMAGING, 2021, 17 (09) : 1059 - 1077
  • [26] Deep Learning and Big DataTechnologies in Medical Image Analysis
    Rastogi, Priyanka
    Singh, Vijendra
    Yadav, Monika
    2018 FIFTH INTERNATIONAL CONFERENCE ON PARALLEL, DISTRIBUTED AND GRID COMPUTING (IEEE PDGC), 2018, : 60 - 63
  • [27] A configurable deep learning framework for medical image analysis
    Jianguo Chen
    Nan Yang
    Mimi Zhou
    Zhaolei Zhang
    Xulei Yang
    Neural Computing and Applications, 2022, 34 : 7375 - 7392
  • [28] Explainable Deep Learning in Spectral and Medical Image Analysis
    Liu, Xuyang
    Duan, Chaoshu
    Cai, Wensheng
    Shao, Xueguang
    PROGRESS IN CHEMISTRY, 2022, 34 (12) : 2561 - 2572
  • [29] An Analysis Of Deep Learning In CXR Medical Image Processing
    Shafi, Syed Mohammed
    Kumar, Sathiya
    JOURNAL OF PHARMACEUTICAL NEGATIVE RESULTS, 2022, 13 : 701 - 709
  • [30] Trends in Deep Learning for Medical Hyperspectral Image Analysis
    Khan, Uzair
    Paheding, Sidike
    Elkin, Colin P.
    Devabhaktuni, Vijaya Kumar
    IEEE ACCESS, 2021, 9 (09): : 79534 - 79548