Siamese tracking combing frequency channel attention with adaptive template

Abstract Siamese network based the tracker is a hot topic in the field of visual object tracking. However, Siamese trackers still have a robustness gap compared with state‐of‐the‐art algorithms. Therefore, focusing on the issue, this letter adds Frequency Channel Attention (FCA) and adaptive templat...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Haibo Pang, Meiqin Xie, Chengming Liu, Rongqi Ma, Linxuan Han
Formato:	article
Lenguaje:	EN
Publicado:	Wiley 2021
Materias:	Telecommunication TK5101-6720
Acceso en línea:	https://doaj.org/article/8490461658e7440a9d3016884e9065f8
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:8490461658e7440a9d3016884e9065f8
record_format	dspace
spelling	oai:doaj.org-article:8490461658e7440a9d3016884e9065f82021-12-01T07:07:54ZSiamese tracking combing frequency channel attention with adaptive template1751-86361751-862810.1049/cmu2.12280https://doaj.org/article/8490461658e7440a9d3016884e9065f82021-12-01T00:00:00Zhttps://doi.org/10.1049/cmu2.12280https://doaj.org/toc/1751-8628https://doaj.org/toc/1751-8636Abstract Siamese network based the tracker is a hot topic in the field of visual object tracking. However, Siamese trackers still have a robustness gap compared with state‐of‐the‐art algorithms. Therefore, focusing on the issue, this letter adds Frequency Channel Attention (FCA) and adaptive template feature map to the framework of Siamese neural network. FCA can enhance feature representation of effective channels and improve feature discrimination by modeling the correlation between each channel of the image. In this algorithm, by theoretical analysis and experimental validation, restriction is broken through a simple yet effective FCA network sampling strategy and a Siamese‐FCA tracker with significant performance gain is successfully trained. Meanwhile, in order to better adjust the proportion between target and background, the tracker selects suitable size of the target feature map. Moreover, extensive ablation studies are conducted to demonstrate the effectiveness of the proposed tracker. Fairly, the experimental results of five test benchmarks, including OTB2013, OTB2015, VOT2016, VOT2018 and UAV123 datasets, shows that the proposed algorithm performs outstanding. In particular, the issue of similarity and small target tracking failure is overcome. The average running frame rate reaches 86 frames per second, which can meet the real‐time requirements.Haibo PangMeiqin XieChengming LiuRongqi MaLinxuan HanWileyarticleTelecommunicationTK5101-6720ENIET Communications, Vol 15, Iss 20, Pp 2493-2502 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Telecommunication TK5101-6720
spellingShingle	Telecommunication TK5101-6720 Haibo Pang Meiqin Xie Chengming Liu Rongqi Ma Linxuan Han Siamese tracking combing frequency channel attention with adaptive template
description	Abstract Siamese network based the tracker is a hot topic in the field of visual object tracking. However, Siamese trackers still have a robustness gap compared with state‐of‐the‐art algorithms. Therefore, focusing on the issue, this letter adds Frequency Channel Attention (FCA) and adaptive template feature map to the framework of Siamese neural network. FCA can enhance feature representation of effective channels and improve feature discrimination by modeling the correlation between each channel of the image. In this algorithm, by theoretical analysis and experimental validation, restriction is broken through a simple yet effective FCA network sampling strategy and a Siamese‐FCA tracker with significant performance gain is successfully trained. Meanwhile, in order to better adjust the proportion between target and background, the tracker selects suitable size of the target feature map. Moreover, extensive ablation studies are conducted to demonstrate the effectiveness of the proposed tracker. Fairly, the experimental results of five test benchmarks, including OTB2013, OTB2015, VOT2016, VOT2018 and UAV123 datasets, shows that the proposed algorithm performs outstanding. In particular, the issue of similarity and small target tracking failure is overcome. The average running frame rate reaches 86 frames per second, which can meet the real‐time requirements.
format	article
author	Haibo Pang Meiqin Xie Chengming Liu Rongqi Ma Linxuan Han
author_facet	Haibo Pang Meiqin Xie Chengming Liu Rongqi Ma Linxuan Han
author_sort	Haibo Pang
title	Siamese tracking combing frequency channel attention with adaptive template
title_short	Siamese tracking combing frequency channel attention with adaptive template
title_full	Siamese tracking combing frequency channel attention with adaptive template
title_fullStr	Siamese tracking combing frequency channel attention with adaptive template
title_full_unstemmed	Siamese tracking combing frequency channel attention with adaptive template
title_sort	siamese tracking combing frequency channel attention with adaptive template
publisher	Wiley
publishDate	2021
url	https://doaj.org/article/8490461658e7440a9d3016884e9065f8
work_keys_str_mv	AT haibopang siamesetrackingcombingfrequencychannelattentionwithadaptivetemplate AT meiqinxie siamesetrackingcombingfrequencychannelattentionwithadaptivetemplate AT chengmingliu siamesetrackingcombingfrequencychannelattentionwithadaptivetemplate AT rongqima siamesetrackingcombingfrequencychannelattentionwithadaptivetemplate AT linxuanhan siamesetrackingcombingfrequencychannelattentionwithadaptivetemplate
_version_	1718405423367716864

Siamese tracking combing frequency channel attention with adaptive template

Ejemplares similares