NL-VTON: a non-local virtual try-on network with feature preserving of body and clothes

Abstract In an image based virtual try-on network, both features of the target clothes and the input human body should be preserved. However, current techniques failed to solve the problems of blurriness on complex clothes details and artifacts on human body occlusion regions at the same time. To ta...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Ze Lin Tan, Jing Bai, Shao Min Zhang, Fei Wei Qin
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2021
Materias:	Medicine R Science Q
Acceso en línea:	https://doaj.org/article/aa4f2f6d7b78419e9ab95489d8acf206
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:aa4f2f6d7b78419e9ab95489d8acf206
record_format	dspace
spelling	oai:doaj.org-article:aa4f2f6d7b78419e9ab95489d8acf2062021-12-02T18:01:47ZNL-VTON: a non-local virtual try-on network with feature preserving of body and clothes10.1038/s41598-021-99406-62045-2322https://doaj.org/article/aa4f2f6d7b78419e9ab95489d8acf2062021-10-01T00:00:00Zhttps://doi.org/10.1038/s41598-021-99406-6https://doaj.org/toc/2045-2322Abstract In an image based virtual try-on network, both features of the target clothes and the input human body should be preserved. However, current techniques failed to solve the problems of blurriness on complex clothes details and artifacts on human body occlusion regions at the same time. To tackle this issue, we propose a non-local virtual try-on network NL-VTON. Considering that convolution is a local operation and limited by its convolution kernel size and rectangular receptive field, which is unsuitable for large size non-rigid transformations of persons and clothes in virtual try-on, we introduce a non-local feature attention module and a grid regularization loss so as to capture detailed features of complex clothes, and design a human body segmentation prediction network to further alleviate the artifacts on occlusion regions. The quantitative and qualitative experiments based on the Zalando dataset demonstrate that our proposed method significantly improves the ability to preserve features of bodies and clothes compared with the state-of-the-art methods.Ze Lin TanJing BaiShao Min ZhangFei Wei QinNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-13 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Medicine R Science Q
spellingShingle	Medicine R Science Q Ze Lin Tan Jing Bai Shao Min Zhang Fei Wei Qin NL-VTON: a non-local virtual try-on network with feature preserving of body and clothes
description	Abstract In an image based virtual try-on network, both features of the target clothes and the input human body should be preserved. However, current techniques failed to solve the problems of blurriness on complex clothes details and artifacts on human body occlusion regions at the same time. To tackle this issue, we propose a non-local virtual try-on network NL-VTON. Considering that convolution is a local operation and limited by its convolution kernel size and rectangular receptive field, which is unsuitable for large size non-rigid transformations of persons and clothes in virtual try-on, we introduce a non-local feature attention module and a grid regularization loss so as to capture detailed features of complex clothes, and design a human body segmentation prediction network to further alleviate the artifacts on occlusion regions. The quantitative and qualitative experiments based on the Zalando dataset demonstrate that our proposed method significantly improves the ability to preserve features of bodies and clothes compared with the state-of-the-art methods.
format	article
author	Ze Lin Tan Jing Bai Shao Min Zhang Fei Wei Qin
author_facet	Ze Lin Tan Jing Bai Shao Min Zhang Fei Wei Qin
author_sort	Ze Lin Tan
title	NL-VTON: a non-local virtual try-on network with feature preserving of body and clothes
title_short	NL-VTON: a non-local virtual try-on network with feature preserving of body and clothes
title_full	NL-VTON: a non-local virtual try-on network with feature preserving of body and clothes
title_fullStr	NL-VTON: a non-local virtual try-on network with feature preserving of body and clothes
title_full_unstemmed	NL-VTON: a non-local virtual try-on network with feature preserving of body and clothes
title_sort	nl-vton: a non-local virtual try-on network with feature preserving of body and clothes
publisher	Nature Portfolio
publishDate	2021
url	https://doaj.org/article/aa4f2f6d7b78419e9ab95489d8acf206
work_keys_str_mv	AT zelintan nlvtonanonlocalvirtualtryonnetworkwithfeaturepreservingofbodyandclothes AT jingbai nlvtonanonlocalvirtualtryonnetworkwithfeaturepreservingofbodyandclothes AT shaominzhang nlvtonanonlocalvirtualtryonnetworkwithfeaturepreservingofbodyandclothes AT feiweiqin nlvtonanonlocalvirtualtryonnetworkwithfeaturepreservingofbodyandclothes
_version_	1718378947816718336

NL-VTON: a non-local virtual try-on network with feature preserving of body and clothes

Ejemplares similares