Self-supervised context-aware correlation filter for robust landmark tracking in liver ultrasound sequences

作者全名："Ma, Lin; Wang, Junjie; Gong, Shu; Lan, Libin; Geng, Li; Wang, Siping; Feng, Xin"

作者地址："[Ma, Lin; Wang, Junjie; Lan, Libin; Wang, Siping; Feng, Xin] Chongqing Univ Technol, Coll Comp Sci & Engn, Chongqing, Peoples R China; [Gong, Shu] Chongqing Med Univ, Dept Gastroenterol, Childrens Hosp, Chongqing 400014, Peoples R China; [Geng, Li] City Univ New York NYCCT, New York, NY USA"

通信作者："Gong, S (通讯作者)，Chongqing Med Univ, Dept Gastroenterol, Childrens Hosp, Chongqing 400014, Peoples R China."

来源：BIOMEDICAL ENGINEERING-BIOMEDIZINISCHE TECHNIK

ESI学科分类：MOLECULAR BIOLOGY & GENETICS

WOS号：WOS:001162395300001

JCR分区：Q4

影响因子：1.3

年份：2024

卷号：　

期号：　

开始页：　

结束页：　

文献类型：Article; Early Access

关键词：self-supervised context-aware correlation filter; liver ultrasound landmark tracking; respiratory motion estimation; image-guided radiation therapy

摘要："Objectives: Respiratory motion-induced displacement of internal organs poses a significant challenge in image-guided radiation therapy, particularly affecting liver landmark tracking accuracy. Methods: Addressing this concern, we propose a self-supervised method for robust landmark tracking in long liver ultrasound sequences. Our approach leverages a Siamese-based context-aware correlation filter network, trained by using the consistency loss between forward tracking and back verification. By effectively utilizing both labeled and unlabeled liver ultrasound images, our model, Siam-CCF , mitigates the impact of speckle noise and artifacts on ultrasonic image tracking by a context-aware correlation filter. Additionally, a fusion strategy for template patch feature helps the tracker to obtain rich appearance information around the point-landmark. Results: Siam-CCF achieves a mean tracking error of 0.79 +/- 0.83 mm at a frame rate of 118.6 fps, exhibiting a superior speed-accuracy trade-off on the public MICCAI 2015 Challenge on Liver Ultrasound Tracking (CLUST2015) 2D dataset. This performance won the 5th place on the CLUST2015 2D point-landmark tracking task. Conclusions: Extensive experiments validate the effectiveness of our proposed approach, establishing it as one of the top-performing techniques on the CLUST2015 online leaderboard at the time of this submission."

基金机构：Key project of the Chongqing Technology Innovation and Application Development [cstc2021jscx-dxwtBX0018]; Natural Science Foundation of Chongqing [CSTB2022NSCQ-MSX0493]; Chongqing Postgraduate Scientific Research Innovation Project [CYS23678]; Action Plan for the High-quality Development of Postgraduate Education of Chongqing University of Technology [gzlcx20233200]; Scientific Research Foundation of Chongqing University of Technology [0103210650]; Youth Project of Science and Technology Research Program of Chongqing Education Commission of China [KJQN202301145]

基金资助正文："This work is supported in part by the Key project of the Chongqing Technology Innovation and Application Development under Grant No. cstc2021jscx-dxwtBX0018, and in part by the Natural Science Foundation of Chongqing under Grant No. CSTB2022NSCQ-MSX0493, and in part by the Chongqing Postgraduate Scientific Research Innovation Project under Grant No.CYS23678, and in part by the Action Plan for the High-quality Development of Postgraduate Education of Chongqing University of Technology under Grant No. gzlcx20233200,and in part by the Scientific Research Foundation of Chongqing University of Technology under Grant No.0103210650, and in part by the Youth Project of Science and Technology Research Program of Chongqing Education Commission of China under Grant No. KJQN202301145."