SKT-MOT and DyTracker: A Multiobject Tracking Dataset and a Dynamic Tracker for Speed Skating Video

4区 计算机科学 Q3 Computer Science
Junwu Wang, Zongmin Li, Yachuan Li, Shaobo Yang, Ben Wang, Hua Li
{"title":"SKT-MOT and DyTracker: A Multiobject Tracking Dataset and a Dynamic Tracker for Speed Skating Video","authors":"Junwu Wang, Zongmin Li, Yachuan Li, Shaobo Yang, Ben Wang, Hua Li","doi":"10.1155/2023/3895703","DOIUrl":null,"url":null,"abstract":"Speed skating serves as a significant application domain for multiobject tracking (MOT), presenting unique challenges such as frequent occlusion, highly similar appearances, and motion blur. To address these challenges, this paper constructs an MOT dataset called SKT-MOT for speed skating and analyzes the shortcomings of existing datasets and methods. Accordingly, we propose a dynamic MOT method called DyTracker. The method builds upon the DeepSORT baseline and enhances three key modules. At the global level, we design the track dynamic management (TDM) algorithm. In the motion branch, a novel metric is proposed to evaluate occlusion and Kalman filter dynamic update (KFDU) is implemented. In the appearance branch, we account for the difference in human posture and propose the feature dynamic selection and updating (FDSU) strategy. This makes our DyTracker flexible and efficient to achieve a multiobject tracking accuracy (MOTA) of 93.70% and identification F1 (IDF1) score of 92.39% on SKT-MOT, which is a significant advantage over existing SOTA methods. To validate the generalization of our proposed module, two dynamic update modules are inserted into other methods and validated on the public dataset MOT17, and the accuracy is generally improved by 0.2%–0.6%.","PeriodicalId":22091,"journal":{"name":"Scientific Programming","volume":"70 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Scientific Programming","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1155/2023/3895703","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"Computer Science","Score":null,"Total":0}
引用次数: 0

Abstract

Speed skating serves as a significant application domain for multiobject tracking (MOT), presenting unique challenges such as frequent occlusion, highly similar appearances, and motion blur. To address these challenges, this paper constructs an MOT dataset called SKT-MOT for speed skating and analyzes the shortcomings of existing datasets and methods. Accordingly, we propose a dynamic MOT method called DyTracker. The method builds upon the DeepSORT baseline and enhances three key modules. At the global level, we design the track dynamic management (TDM) algorithm. In the motion branch, a novel metric is proposed to evaluate occlusion and Kalman filter dynamic update (KFDU) is implemented. In the appearance branch, we account for the difference in human posture and propose the feature dynamic selection and updating (FDSU) strategy. This makes our DyTracker flexible and efficient to achieve a multiobject tracking accuracy (MOTA) of 93.70% and identification F1 (IDF1) score of 92.39% on SKT-MOT, which is a significant advantage over existing SOTA methods. To validate the generalization of our proposed module, two dynamic update modules are inserted into other methods and validated on the public dataset MOT17, and the accuracy is generally improved by 0.2%–0.6%.
SKT-MOT和DyTracker:一个多目标跟踪数据集和一个用于速滑视频的动态跟踪器
速滑作为多目标跟踪(MOT)的重要应用领域,呈现出诸如频繁遮挡、高度相似外观和运动模糊等独特挑战。为了解决这些问题,本文构建了一个速度滑冰的MOT数据集SKT-MOT,并分析了现有数据集和方法的不足。因此,我们提出了一种动态MOT方法,称为DyTracker。该方法以DeepSORT基线为基础,增强了三个关键模块。在全局层面,设计了航迹动态管理(TDM)算法。在运动分支中,提出了一种新的遮挡评价度量,并实现了卡尔曼滤波动态更新(KFDU)。在外观分支中,考虑了人体姿态的差异,提出了特征动态选择与更新策略。这使得我们的DyTracker灵活高效,在SKT-MOT上实现了93.70%的多目标跟踪精度(MOTA)和92.39%的识别F1 (IDF1)分数,与现有的SOTA方法相比具有显著的优势。为了验证我们提出的模块的泛化性,我们将两个动态更新模块插入到其他方法中,并在公共数据集MOT17上进行验证,准确率一般提高0.2%-0.6%。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Scientific Programming
Scientific Programming 工程技术-计算机:软件工程
自引率
0.00%
发文量
1059
审稿时长
>12 weeks
期刊介绍: Scientific Programming is a peer-reviewed, open access journal that provides a meeting ground for research results in, and practical experience with, software engineering environments, tools, languages, and models of computation aimed specifically at supporting scientific and engineering computing. The journal publishes papers on language, compiler, and programming environment issues for scientific computing. Of particular interest are contributions to programming and software engineering for grid computing, high performance computing, processing very large data sets, supercomputing, visualization, and parallel computing. All languages used in scientific programming as well as scientific programming libraries are within the scope of the journal.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信