SDM-NET: Deep Generative Network for Structured Deformable Mesh

doi:10.1145/3355089.3356488

	SDM-NET: Deep Generative Network for Structured Deformable Mesh
	Gao, Lin 1,6; Yang, Jie 1,2,6; Wu, Tong 1,2,6; Yuan, Yu-Jie 1,2,6; Fu, Hongbo 3; Lai, Yu-Kun 4; Zhang, Hao 5
	2019-11-01
发表期刊	ACM TRANSACTIONS ON GRAPHICS
ISSN	0730-0301
卷号	38 期号:6 页码:15
摘要	We introduce SDM-NET, a deep generative neural network which produces structured deformable meshes. Specifically, the network is trained to generate a spatial arrangement of closed, deformable mesh parts, which respects the global part structure of a shape collection, e.g., chairs, airplanes, etc. Our key observation is that while the overall structure of a 3D shape can be complex, the shape can usually be decomposed into a set of parts, each homeomorphic to a box, and the finer-scale geometry of the part can be recovered by deforming the box. The architecture of SDM-NET is that of a two-level variational autoencoder (VAE). At the part level, a PartVAE learns a deformable model of part geometries. At the structural level, we train a Structured Parts VAE (SP-VAE), which jointly learns the part structure of a shape collection and the part geometries, ensuring the coherence between global shape structure and surface details. Through extensive experiments and comparisons with the state-of-the-art deep generative models of shapes, we demonstrate the superiority of SDM-NET in generating meshes with visual quality, flexible topology, and meaningful structures, benefiting shape interpolation and other subsequent modeling tasks.
关键词	Shape representation variational autoencoder structure geometric details generation
DOI	10.1145/3355089.3356488
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[61828204] ; National Natural Science Foundation of China[61872440] ; Beijing Municipal Natural Science Foundation[L182016] ; Youth Innovation Promotion Association CAS ; CCF-Tencent Open Fund ; Research Grants Council of the Hong Kong Special Administrative Region, China[CityU 11237116] ; Research Grants Council of the Hong Kong Special Administrative Region, China[CityU 11300615] ; Centre for Applied Computing and Interactive Media (ACIM) of School of Creative Media, CityU ; Sense Time Research Fund
WOS研究方向	Computer Science
WOS类目	Computer Science, Software Engineering
WOS记录号	WOS:000498397300091
出版者	ASSOC COMPUTING MACHINERY
引用统计	被引频次：128[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/14939
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Gao, Lin
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Beijing, Peoples R China 2.Univ Chinese Acad Sci, Beijing, Peoples R China 3.City Univ Hong Kong, Sch Creat Media, Hong Kong, Peoples R China 4.Cardiff Univ, Sch Comp Sci & Informat, Cardiff, S Glam, Wales 5.Simon Fraser Univ, Sch Comp Sci, Burnaby, BC, Canada 6.Chinese Acad Sci, Inst Comp Technol, Beijing Key Lab Mobile Comp & Pervas Device, Beijing, Peoples R China
推荐引用方式 GB/T 7714	Gao, Lin,Yang, Jie,Wu, Tong,et al. SDM-NET: Deep Generative Network for Structured Deformable Mesh[J]. ACM TRANSACTIONS ON GRAPHICS,2019,38(6):15.
APA	Gao, Lin.,Yang, Jie.,Wu, Tong.,Yuan, Yu-Jie.,Fu, Hongbo.,...&Zhang, Hao.(2019).SDM-NET: Deep Generative Network for Structured Deformable Mesh.ACM TRANSACTIONS ON GRAPHICS,38(6),15.
MLA	Gao, Lin,et al."SDM-NET: Deep Generative Network for Structured Deformable Mesh".ACM TRANSACTIONS ON GRAPHICS 38.6(2019):15.