《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (6): 1930-1937.DOI: 10.11772/j.issn.1001-9081.2022050674

• 多媒体计算与计算机仿真 • 上一篇    下一篇

基于实例分割与毕达哥拉斯模糊决策的目标跟踪

赵元龙1, 单玉刚2(), 袁杰1, 赵康迪1   

  1. 1.新疆大学 电气工程学院,乌鲁木齐 830017
    2.湖北文理学院 教育学院,湖北 襄阳 441053
  • 收稿日期:2022-05-11 修回日期:2023-02-15 接受日期:2023-02-15 发布日期:2023-06-08 出版日期:2023-06-10
  • 通讯作者: 单玉刚
  • 作者简介:赵元龙(1994—),男,湖北襄阳人,硕士研究生,主要研究方向:目标跟踪、目标检测
    单玉刚(1971—),男,辽宁沈阳人,讲师,博士,主要研究方向:模式识别Email:shanyg@21cn.com
    袁杰(1975—),男,新疆乌鲁木齐人,副教授,博士,主要研究方向:机器人控制
    赵康迪(1995—),男,河南南阳人,硕士研究生,主要研究方向:目标检测。
  • 基金资助:
    国家自然科学基金资助项目(62263031);新疆维吾尔自治区自然科学基金资助项目(2022D01C53);湖北省教育科学规划基金资助项目(2021GA048);教育部产学合作协同育人项目(202102602033);襄阳市科技计划项目(高新领域)(2020ABH001799);湖北文理学院“双百行动计划”项目(PYSB20202016);湖北文理学院2021年度教师科研能力培育基金资助项目(2021KPGPSK08)

Object tracking based on instance segmentation and Pythagorean fuzzy decision-making

Yuanlong ZHAO1, Yugang SHAN2(), Jie YUAN1, Kangdi ZHAO1   

  1. 1.School of Electrical Engineering,Xinjiang University,Urumqi Xinjiang 830017,China
    2.School of Education,Hubei University of Arts and Science,Xiangyang Hubei 441053,China
  • Received:2022-05-11 Revised:2023-02-15 Accepted:2023-02-15 Online:2023-06-08 Published:2023-06-10
  • Contact: Yugang SHAN
  • About author:ZHAO Yuanlong, born in 1994, M. S. candidate. His research interests include object tracking, object detection.
    YUAN Jie, born in 1975, Ph. D., associate professor. His research interests include robot control.
    ZHAO Kangdi, born in 1995, M. S. candidate. His research interests include object detection.
  • Supported by:
    National Natural Science Foundation of China(62263031);Natural Science Foundation of Xinjiang Uygur Autonomous Region(2022D01C53);Education Science Planning Foundation of Hubei(2021GA048);Industry-University Cooperative Education Project of Ministry of Education(202102602033);Science and Technology Program of Xiangyang (High-Tech Field)(2020ABH001799);“Double Hundred Action Plan” Project of Hubei University of Arts and Science(PYSB20202016);2021 Teachers’ Research Ability Training Fund of Hubei University of Arts and Science(2021KPGPSK08)

摘要:

为了解决目标跟踪中的尺度变化、相似性干扰、遮挡等问题,提出一种基于实例分割与毕达哥拉斯模糊决策的目标跟踪算法。在实例分割网络YOLACT++ (improved You Only Look At CoefficienTs)的基础上,融合3种不同的匹配方式针对不同场景预测跟踪结果;同时提出一种基于毕达哥拉斯模糊决策的模板更新机制,即根据预测结果的质量作出是否更新目标模板和更换匹配方式的决定。实验结果表明,所提算法能够更准确地跟踪存在尺度变化、相似性干扰、遮挡等问题的视频序列。相较于SiamMask算法,所提算法在DAVIS 2016、DAVIS 2017数据集上的区域相似度分别提高了12.3、15.3个百分点,在VOT2016、VOT2018数据集上的预期平均重叠率(EAO)分别提高了4.2、4.1个百分点,且所提算法的平均跟踪速度为每秒32.00帧,满足实时性要求。

关键词: 图像处理, 目标跟踪, 毕达哥拉斯模糊决策, 实例分割, 孪生网络

Abstract:

In order to solve the problems of scale change, similarity interference and occlusion in object tracking, an object tracking algorithm based on instance segmentation and Pythagorean fuzzy decision-making was proposed. Based on the instance segmentation network YOLACT++ (improved You Only Look At CoefficienTs), three different matching methods were integrated to predict the tracking results for different scenes. At the same time, a template update mechanism based on Pythagorean fuzzy decision-making was proposed by which whether to update the object template and replace the matching method was determined according to the quality of the prediction results. Experimental results show that the proposed algorithm can track the video sequences with scale change, similarity interference, occlusion and other problems more accurately. Compared with SiamMask algorithm, the proposed algorithm has the regional similarity on DAVIS 2016 and DAVIS 2017 datasets increased by 12.3 and 15.3 percentage points, respectively, and the Expected Average Overlap rate (EAO) on VOT2016 and VOT2018 datasets increased by 4.2 and 4.1 percentage points, respectively. Meanwhile, the average tracking speed of the proposed algorithm is 32.00 frames per second, meeting real-time requirements.

Key words: image processing, object tracking, Pythagorean fuzzy decision-making, instance segmentation, Siamese network

中图分类号: