VOVTrack: Exploring the Potentiality in Videos for Open-Vocabulary Object Tracking

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Qian, Zekun, Han, Ruize, Hou, Junhui, Song, Linqi, Feng, Wei
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912069021859840
author Qian, Zekun
Han, Ruize
Hou, Junhui
Song, Linqi
Feng, Wei
author_facet Qian, Zekun
Han, Ruize
Hou, Junhui
Song, Linqi
Feng, Wei
contents Open-vocabulary multi-object tracking (OVMOT) represents a critical new challenge involving the detection and tracking of diverse object categories in videos, encompassing both seen categories (base classes) and unseen categories (novel classes). This issue amalgamates the complexities of open-vocabulary object detection (OVD) and multi-object tracking (MOT). Existing approaches to OVMOT often merge OVD and MOT methodologies as separate modules, predominantly focusing on the problem through an image-centric lens. In this paper, we propose VOVTrack, a novel method that integrates object states relevant to MOT and video-centric training to address this challenge from a video object tracking standpoint. First, we consider the tracking-related state of the objects during tracking and propose a new prompt-guided attention mechanism for more accurate localization and classification (detection) of the time-varying objects. Subsequently, we leverage raw video data without annotations for training by formulating a self-supervised object similarity learning technique to facilitate temporal object association (tracking). Experimental results underscore that VOVTrack outperforms existing methods, establishing itself as a state-of-the-art solution for open-vocabulary tracking task.
format Preprint
id arxiv_https___arxiv_org_abs_2410_08529
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle VOVTrack: Exploring the Potentiality in Videos for Open-Vocabulary Object Tracking
Qian, Zekun
Han, Ruize
Hou, Junhui
Song, Linqi
Feng, Wei
Computer Vision and Pattern Recognition
Artificial Intelligence
Open-vocabulary multi-object tracking (OVMOT) represents a critical new challenge involving the detection and tracking of diverse object categories in videos, encompassing both seen categories (base classes) and unseen categories (novel classes). This issue amalgamates the complexities of open-vocabulary object detection (OVD) and multi-object tracking (MOT). Existing approaches to OVMOT often merge OVD and MOT methodologies as separate modules, predominantly focusing on the problem through an image-centric lens. In this paper, we propose VOVTrack, a novel method that integrates object states relevant to MOT and video-centric training to address this challenge from a video object tracking standpoint. First, we consider the tracking-related state of the objects during tracking and propose a new prompt-guided attention mechanism for more accurate localization and classification (detection) of the time-varying objects. Subsequently, we leverage raw video data without annotations for training by formulating a self-supervised object similarity learning technique to facilitate temporal object association (tracking). Experimental results underscore that VOVTrack outperforms existing methods, establishing itself as a state-of-the-art solution for open-vocabulary tracking task.
title VOVTrack: Exploring the Potentiality in Videos for Open-Vocabulary Object Tracking
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2410.08529