Anchor-Free Tracker Based on Space-Time Memory Network

Guang HAN; Chen CAO; Jixin LIU; Sam KWONG

doi:10.1109/MMUL.2022.3207016

Anchor-Free Tracker Based on Space-Time Memory Network

Guang HAN, Chen CAO, Jixin LIU, Sam KWONG

Research output: Journal Publications › Journal Article (refereed) › peer-review

Abstract

In the visual object tracking task, the existing trackers cannot well solve the appearance deformation, occlusion, and similar object interference, etc. To address these problems, this article proposes a new Anchor-free Tracker based on Space-time Memory Network (ATSMN). In this work, we innovatively use the space-time memory network, memory feature fusion network, and transformer feature cross fusion network. Through the synergy of above-mentioned innovations, tracker can make full use of temporal context information in the memory frames related to the object and better adapt to the appearance change of the object, which can obtain accurate classification and regression results. Extensive experimental results on challenging benchmarks show that ATSMN can achieve the SOTA level tracking performance compared with other advanced trackers.

Original language	English
Number of pages	9
Journal	IEEE Multimedia
DOIs	https://doi.org/10.1109/MMUL.2022.3207016
Publication status	E-pub ahead of print - 23 Sept 2022
Externally published	Yes

Bibliographical note

This work was supported in part by the Natural Science Foundation of China NSFC under Grants 61871445 and 61302156, and in part by Key R and D Foundation Project of Jiangsu Province under Grant BE2016001-4.

Keywords

Anchor-free
Feature cross fusion
Object tracking
Space-time memory network

Access to Document

10.1109/MMUL.2022.3207016

Cite this

@article{c0deaf29848342b589bc6938893e2a95,

title = "Anchor-Free Tracker Based on Space-Time Memory Network",

abstract = "In the visual object tracking task, the existing trackers cannot well solve the appearance deformation, occlusion, and similar object interference, etc. To address these problems, this article proposes a new Anchor-free Tracker based on Space-time Memory Network (ATSMN). In this work, we innovatively use the space-time memory network, memory feature fusion network, and transformer feature cross fusion network. Through the synergy of above-mentioned innovations, tracker can make full use of temporal context information in the memory frames related to the object and better adapt to the appearance change of the object, which can obtain accurate classification and regression results. Extensive experimental results on challenging benchmarks show that ATSMN can achieve the SOTA level tracking performance compared with other advanced trackers.",

keywords = "Anchor-free, Feature cross fusion, Object tracking, Space-time memory network",

author = "Guang HAN and Chen CAO and Jixin LIU and Sam KWONG",

note = "This work was supported in part by the Natural Science Foundation of China NSFC under Grants 61871445 and 61302156, and in part by Key R and D Foundation Project of Jiangsu Province under Grant BE2016001-4.",

year = "2022",

month = sep,

day = "23",

doi = "10.1109/MMUL.2022.3207016",

language = "English",

journal = "IEEE Multimedia",

issn = "1070-986X",

publisher = "IEEE Computer Society",

}

TY - JOUR

T1 - Anchor-Free Tracker Based on Space-Time Memory Network

AU - HAN, Guang

AU - CAO, Chen

AU - LIU, Jixin

AU - KWONG, Sam

N1 - This work was supported in part by the Natural Science Foundation of China NSFC under Grants 61871445 and 61302156, and in part by Key R and D Foundation Project of Jiangsu Province under Grant BE2016001-4.

PY - 2022/9/23

Y1 - 2022/9/23

N2 - In the visual object tracking task, the existing trackers cannot well solve the appearance deformation, occlusion, and similar object interference, etc. To address these problems, this article proposes a new Anchor-free Tracker based on Space-time Memory Network (ATSMN). In this work, we innovatively use the space-time memory network, memory feature fusion network, and transformer feature cross fusion network. Through the synergy of above-mentioned innovations, tracker can make full use of temporal context information in the memory frames related to the object and better adapt to the appearance change of the object, which can obtain accurate classification and regression results. Extensive experimental results on challenging benchmarks show that ATSMN can achieve the SOTA level tracking performance compared with other advanced trackers.

AB - In the visual object tracking task, the existing trackers cannot well solve the appearance deformation, occlusion, and similar object interference, etc. To address these problems, this article proposes a new Anchor-free Tracker based on Space-time Memory Network (ATSMN). In this work, we innovatively use the space-time memory network, memory feature fusion network, and transformer feature cross fusion network. Through the synergy of above-mentioned innovations, tracker can make full use of temporal context information in the memory frames related to the object and better adapt to the appearance change of the object, which can obtain accurate classification and regression results. Extensive experimental results on challenging benchmarks show that ATSMN can achieve the SOTA level tracking performance compared with other advanced trackers.

KW - Anchor-free

KW - Feature cross fusion

KW - Object tracking

KW - Space-time memory network

UR - http://www.scopus.com/inward/record.url?scp=85139427165&partnerID=8YFLogxK

U2 - 10.1109/MMUL.2022.3207016

DO - 10.1109/MMUL.2022.3207016

M3 - Journal Article (refereed)

SN - 1070-986X

JO - IEEE Multimedia

JF - IEEE Multimedia

ER -

Anchor-Free Tracker Based on Space-Time Memory Network

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this