Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Huang, Junchao, Wu, Xiaoqi He Yebo, Zhao, Sheng
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2307.14591
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910904575066112
author Huang, Junchao
Wu, Xiaoqi He Yebo
Zhao, Sheng
author_facet Huang, Junchao
Wu, Xiaoqi He Yebo
Zhao, Sheng
contents The purpose of multi-object tracking (MOT) is to continuously track and identify objects detected in videos. Currently, most methods for multi-object tracking model the motion information and combine it with appearance information to determine and track objects. In this paper, unfalsified control is employed to address the ID-switch problem in multi-object tracking. We establish sequences of appearance information variations for the trajectories during the tracking process and design a detection and rectification module specifically for ID-switch detection and recovery. We also propose a simple and effective strategy to address the issue of ambiguous matching of appearance information during the data association process. Experimental results on publicly available MOT datasets demonstrate that the tracker exhibits excellent effectiveness and robustness in handling tracking errors caused by occlusions and rapid movements.
format Preprint
id arxiv_https___arxiv_org_abs_2307_14591
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle The detection and rectification for identity-switch based on unfalsified control
Huang, Junchao
Wu, Xiaoqi He Yebo
Zhao, Sheng
Computer Vision and Pattern Recognition
Artificial Intelligence
The purpose of multi-object tracking (MOT) is to continuously track and identify objects detected in videos. Currently, most methods for multi-object tracking model the motion information and combine it with appearance information to determine and track objects. In this paper, unfalsified control is employed to address the ID-switch problem in multi-object tracking. We establish sequences of appearance information variations for the trajectories during the tracking process and design a detection and rectification module specifically for ID-switch detection and recovery. We also propose a simple and effective strategy to address the issue of ambiguous matching of appearance information during the data association process. Experimental results on publicly available MOT datasets demonstrate that the tracker exhibits excellent effectiveness and robustness in handling tracking errors caused by occlusions and rapid movements.
title The detection and rectification for identity-switch based on unfalsified control
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2307.14591