Masked Spectral Bands Modeling With Shifted Windows: An Excellent Self-Supervised Learner for Classification of Medical Hyperspectral Images

作者全名："Li, Yuan; Wu, Ruoyu; Tan, Qijuan; Yang, Zhengchun; Huang, Hong"

作者地址："[Li, Yuan; Wu, Ruoyu; Huang, Hong] Chongqing Univ, Key Lab Optoelect Technol & Syst Educ, Minist China, Chongqing 400044, Peoples R China; [Tan, Qijuan] Chongqing Univ, Dept Radiol, Canc Hosp, Chongqing 400030, Peoples R China; [Tan, Qijuan] Chongqing Canc Inst, Chongqing 400030, Peoples R China; [Tan, Qijuan] Chongqing Canc Hosp, Chongqing 400030, Peoples R China; [Yang, Zhengchun] Chongqing Med Univ, Women & Childrens Hosp, Chongqing 401147, Peoples R China"

通信作者："Tan, QJ (通讯作者)，Chongqing Univ, Dept Radiol, Canc Hosp, Chongqing 400030, Peoples R China.; Tan, QJ (通讯作者)，Chongqing Canc Inst, Chongqing 400030, Peoples R China.; Tan, QJ (通讯作者)，Chongqing Canc Hosp, Chongqing 400030, Peoples R China.; Yang, ZC (通讯作者)，Chongqing Med Univ, Women & Childrens Hosp, Chongqing 401147, Peoples R China."

来源：IEEE SIGNAL PROCESSING LETTERS

ESI学科分类：ENGINEERING

WOS号：WOS:000988581400004

JCR分区：Q2

影响因子：3.2

年份：2023

卷号：30

期号：　

开始页：543

结束页：547

文献类型：Article

关键词：Hyperspectral imaging; Feature extraction; Transformers; Training; Solid modeling; Signal processing algorithms; Medical diagnostic imaging; Medical hyperspectral images; classification; self-supervised learning; transformer; shift windows

摘要："Hyperspectral imaging has become a popular imaging technique in the medical field, and the development of algorithms for computer-aided diagnosis (CAD) is urgently required. Traditional deep learning techniques require a lot of annotated data, which is a burden on doctors. Self-supervised learning (SSL) is a solution for extracting feature representations from unlabeled data. However, traditional CNN-based SSL algorithms cannot explore relations between neighboring and long-range spectral bands, which limits classification performance. In this letter, the proposed solution is a novel SSL method using a transformer-based technique called masked spectral bands modeling with shifted windows (MSBMSW). This method predicts masked spectral bands as the pretext task and uses a self-attention mechanism with shifted windows to capture the divergence of neighboring spectral bands and enhance information exchange between long-range spectral bands. Experimental results demonstrate that MSBMSW achieves better classification results than many state-of-the-art methods and has potential clinical value for CAD of MHSIs."

基金机构：National Natural Science Foundation of China [42071302]; Innovation Program for Chongqing Overseas Returnees [cx2019144]; Graduate Research and Innovation Foundation of Chongqing [CYB21060]; Higher Education and Research Grants of NVIDIA

基金资助正文："This work was supported in part by the National Natural Science Foundation of China under Grant 42071302, in part by the Innovation Program for Chongqing Overseas Returnees under Grant cx2019144, in part by the Graduate Research and Innovation Foundation of Chongqing under Grant CYB21060, and in part by the Higher Education and Research Grants of NVIDIA. The associate editor coordinating the review of this manuscript and approving it for publication was Dr. Aiping Liu.(Corresponding authors: Qijuan Tan; Zhengchun Yang; HongHuang.)."