一体化多目标跟踪算法研究综述  

A Survey on One-Shot Multi-Object Tracking Algorithm

作　　者：周雪[1,2] 梁超[2] 何均洋 唐瀚林 ZHOU Xue;LIANG Chao;HE Junyang;TANG Hanlin(Shenzhen Institute for Advanced Study,University of Electronic Science and Technology of China Shenzhen Guangdong 518110;School of Automation Engineering,University of Electronic Science and Technology of China Chengdu 611731)

机构地区：[1]电子科技大学(深圳)高等研究院,广东深圳518110 [2]电子科技大学自动化工程学院,成都611731

出　　处：《电子科技大学学报》2022年第5期728-736,共9页Journal of University of Electronic Science and Technology of China

基　　金：国家自然科学基金(61972071,U20A20184);四川省科技计划(2020YJ0036);厅市共建智能终端四川省重点实验室开放课题(SCITLAB-1005)。

摘　　要：视觉多目标跟踪算法(MOT)一直是计算机视觉与视频图像智能分析领域的一个研究热点。近年来,随着深度学习的发展及实际应用需要,越来越多性能优异的一体化多目标跟踪算法被提出,受到研究者的青睐。对近年来广受关注的一体化多目标跟踪算法进行了系统性的综述。从不同的一体化构建思路出发,梳理包括构建出发点、框架设计、方法优缺点、研究趋势等方面的内容,并在权威的MOT Challenge系列数据集上进行性能比较,定量地分析不同的一体化方法的优势和局限性。最后,结合研究现状,提出了一体化多目标跟踪需要重点关注的若干问题及未来展望。Visual multiple object tracking(MOT)has become a hot issue in computer vision and intelligent analysis of video images.In recent years,with the development of deep learning and practical application needs,more and more one-shot MOT algorithms with outstanding performance have been proposed,attracting much attention from researchers.This paper systematically reviews the popular one-shot MOT algorithms.From different construction ideas,the paper summarizes the motivation,framework design,strengths and weaknesses of methods,research trends,etc.Afterwards,we compare the performances of the one-shot MOT algorithms on the public testing set MOT Challenge,and quantitatively analyze the advantages and limitations of different one-shot methods.Finally,some future thoughts,foresight,and key issues that need to be focused on are introduced based on the research status.

关 键 词：数据关联 多目标跟踪 目标检测 一体化深度神经网络 单目标跟踪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]