Siamese anchor-free object tracking with multiscale spatial attentions

Abstract Recently, object trackers based on Siamese networks have attracted considerable attentions due to their remarkable tracking performance and widespread application. Especially, the anchor-based methods exploit the region proposal subnetwork to get accurate prediction of a target and make gre...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Jianming Zhang, Benben Huang, Zi Ye, Li-Dan Kuang, Xin Ning
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2021
Materias:	Medicine R Science Q
Acceso en línea:	https://doaj.org/article/b32f336b5f24410d8b758f349917043a
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:b32f336b5f24410d8b758f349917043a
record_format	dspace
spelling	oai:doaj.org-article:b32f336b5f24410d8b758f349917043a2021-11-28T12:19:32ZSiamese anchor-free object tracking with multiscale spatial attentions10.1038/s41598-021-02095-42045-2322https://doaj.org/article/b32f336b5f24410d8b758f349917043a2021-11-01T00:00:00Zhttps://doi.org/10.1038/s41598-021-02095-4https://doaj.org/toc/2045-2322Abstract Recently, object trackers based on Siamese networks have attracted considerable attentions due to their remarkable tracking performance and widespread application. Especially, the anchor-based methods exploit the region proposal subnetwork to get accurate prediction of a target and make great performance improvement. However, those trackers cannot capture the spatial information very well and the pre-defined anchors will hinder robustness. To solve these problems, we propose a Siamese-based anchor-free object tracking algorithm with multiscale spatial attentions in this paper. Firstly, we take ResNet-50 as the backbone network to generate multiscale features of both template patch and search regions. Secondly, we propose the spatial attention extraction (SAE) block to capture the spatial information among all positions in the template and search region feature maps. Thirdly, we put these features into the SAE block to get the multiscale spatial attentions. Finally, an anchor-free classification and regression subnetwork is used for predicting the location of the target. Unlike anchor-based methods, our tracker directly predicts the target position without predefined parameters. Extensive experiments with state-of-the-art trackers are carried out on four challenging visual object tracking benchmarks: OTB100, UAV123, VOT2016 and GOT-10k. Those experimental results confirm the effectiveness of our proposed tracker.Jianming ZhangBenben HuangZi YeLi-Dan KuangXin NingNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-14 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Medicine R Science Q
spellingShingle	Medicine R Science Q Jianming Zhang Benben Huang Zi Ye Li-Dan Kuang Xin Ning Siamese anchor-free object tracking with multiscale spatial attentions
description	Abstract Recently, object trackers based on Siamese networks have attracted considerable attentions due to their remarkable tracking performance and widespread application. Especially, the anchor-based methods exploit the region proposal subnetwork to get accurate prediction of a target and make great performance improvement. However, those trackers cannot capture the spatial information very well and the pre-defined anchors will hinder robustness. To solve these problems, we propose a Siamese-based anchor-free object tracking algorithm with multiscale spatial attentions in this paper. Firstly, we take ResNet-50 as the backbone network to generate multiscale features of both template patch and search regions. Secondly, we propose the spatial attention extraction (SAE) block to capture the spatial information among all positions in the template and search region feature maps. Thirdly, we put these features into the SAE block to get the multiscale spatial attentions. Finally, an anchor-free classification and regression subnetwork is used for predicting the location of the target. Unlike anchor-based methods, our tracker directly predicts the target position without predefined parameters. Extensive experiments with state-of-the-art trackers are carried out on four challenging visual object tracking benchmarks: OTB100, UAV123, VOT2016 and GOT-10k. Those experimental results confirm the effectiveness of our proposed tracker.
format	article
author	Jianming Zhang Benben Huang Zi Ye Li-Dan Kuang Xin Ning
author_facet	Jianming Zhang Benben Huang Zi Ye Li-Dan Kuang Xin Ning
author_sort	Jianming Zhang
title	Siamese anchor-free object tracking with multiscale spatial attentions
title_short	Siamese anchor-free object tracking with multiscale spatial attentions
title_full	Siamese anchor-free object tracking with multiscale spatial attentions
title_fullStr	Siamese anchor-free object tracking with multiscale spatial attentions
title_full_unstemmed	Siamese anchor-free object tracking with multiscale spatial attentions
title_sort	siamese anchor-free object tracking with multiscale spatial attentions
publisher	Nature Portfolio
publishDate	2021
url	https://doaj.org/article/b32f336b5f24410d8b758f349917043a
work_keys_str_mv	AT jianmingzhang siameseanchorfreeobjecttrackingwithmultiscalespatialattentions AT benbenhuang siameseanchorfreeobjecttrackingwithmultiscalespatialattentions AT ziye siameseanchorfreeobjecttrackingwithmultiscalespatialattentions AT lidankuang siameseanchorfreeobjecttrackingwithmultiscalespatialattentions AT xinning siameseanchorfreeobjecttrackingwithmultiscalespatialattentions
_version_	1718408097391706112

Siamese anchor-free object tracking with multiscale spatial attentions

Ejemplares similares