MV-LFN: Multi-view based local information fusion network for 3D shape recognition

3D shape recognition is a challenging task due to the difficulty of representing the complex structure of 3D shapes. Recently, the view-based approaches that utilize the multiple views rendered from the shape for visual information extraction and feature aggregation to generate a global shape descri...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Jing Zhang, Dangdang Zhou, Yue Zhao, Weizhi Nie, Yuting Su
Formato:	article
Lenguaje:	EN
Publicado:	Elsevier 2021
Materias:	3D model retrieval Multi-view Local information Information technology T58.5-58.64
Acceso en línea:	https://doaj.org/article/6f805698b6b544f894e4cfa2758fdc58
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:6f805698b6b544f894e4cfa2758fdc58
record_format	dspace
spelling	oai:doaj.org-article:6f805698b6b544f894e4cfa2758fdc582021-11-06T04:33:12ZMV-LFN: Multi-view based local information fusion network for 3D shape recognition2468-502X10.1016/j.visinf.2021.09.003https://doaj.org/article/6f805698b6b544f894e4cfa2758fdc582021-09-01T00:00:00Zhttp://www.sciencedirect.com/science/article/pii/S2468502X21000395https://doaj.org/toc/2468-502X3D shape recognition is a challenging task due to the difficulty of representing the complex structure of 3D shapes. Recently, the view-based approaches that utilize the multiple views rendered from the shape for visual information extraction and feature aggregation to generate a global shape descriptor, achieved promising performance. However, the view-based approaches commonly ignore the exploration and utilization of local information in the multiple views, which influences the effectiveness of generated features. In this paper, we design a novel Multi-view based Local Information Fusion Network (MV-LFN) for the 3D shape recognition task. The local correlation attention mechanism (LCAM) is introduced to exploit the local correlations in the feature maps for generating a more effective view descriptor. Then, we hierarchically aggregate the multi-view feature maps to generate a shape super matrix (SSM). The local information is effectively extracted and maintained during the multi-view aggregation process, and the discrimination of shape descriptors is significantly improved. We conduct comparative experiments on the ModelNet and ShapeNetCore55 databases. The experimental performances effectively validate the superiority of MV-LFN.Jing ZhangDangdang ZhouYue ZhaoWeizhi NieYuting SuElsevierarticle3D model retrievalMulti-viewLocal informationInformation technologyT58.5-58.64ENVisual Informatics, Vol 5, Iss 3, Pp 114-119 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	3D model retrieval Multi-view Local information Information technology T58.5-58.64
spellingShingle	3D model retrieval Multi-view Local information Information technology T58.5-58.64 Jing Zhang Dangdang Zhou Yue Zhao Weizhi Nie Yuting Su MV-LFN: Multi-view based local information fusion network for 3D shape recognition
description	3D shape recognition is a challenging task due to the difficulty of representing the complex structure of 3D shapes. Recently, the view-based approaches that utilize the multiple views rendered from the shape for visual information extraction and feature aggregation to generate a global shape descriptor, achieved promising performance. However, the view-based approaches commonly ignore the exploration and utilization of local information in the multiple views, which influences the effectiveness of generated features. In this paper, we design a novel Multi-view based Local Information Fusion Network (MV-LFN) for the 3D shape recognition task. The local correlation attention mechanism (LCAM) is introduced to exploit the local correlations in the feature maps for generating a more effective view descriptor. Then, we hierarchically aggregate the multi-view feature maps to generate a shape super matrix (SSM). The local information is effectively extracted and maintained during the multi-view aggregation process, and the discrimination of shape descriptors is significantly improved. We conduct comparative experiments on the ModelNet and ShapeNetCore55 databases. The experimental performances effectively validate the superiority of MV-LFN.
format	article
author	Jing Zhang Dangdang Zhou Yue Zhao Weizhi Nie Yuting Su
author_facet	Jing Zhang Dangdang Zhou Yue Zhao Weizhi Nie Yuting Su
author_sort	Jing Zhang
title	MV-LFN: Multi-view based local information fusion network for 3D shape recognition
title_short	MV-LFN: Multi-view based local information fusion network for 3D shape recognition
title_full	MV-LFN: Multi-view based local information fusion network for 3D shape recognition
title_fullStr	MV-LFN: Multi-view based local information fusion network for 3D shape recognition
title_full_unstemmed	MV-LFN: Multi-view based local information fusion network for 3D shape recognition
title_sort	mv-lfn: multi-view based local information fusion network for 3d shape recognition
publisher	Elsevier
publishDate	2021
url	https://doaj.org/article/6f805698b6b544f894e4cfa2758fdc58
work_keys_str_mv	AT jingzhang mvlfnmultiviewbasedlocalinformationfusionnetworkfor3dshaperecognition AT dangdangzhou mvlfnmultiviewbasedlocalinformationfusionnetworkfor3dshaperecognition AT yuezhao mvlfnmultiviewbasedlocalinformationfusionnetworkfor3dshaperecognition AT weizhinie mvlfnmultiviewbasedlocalinformationfusionnetworkfor3dshaperecognition AT yutingsu mvlfnmultiviewbasedlocalinformationfusionnetworkfor3dshaperecognition
_version_	1718443858908413952

MV-LFN: Multi-view based local information fusion network for 3D shape recognition

Ejemplares similares