DeePSD: Automatic Deep Skinning And Pose Space Deformation For 3D Garment Animation

被引:18
作者
Bertiche, Hugo [1 ,2 ]
Madadi, Meysam [1 ]
Tylson, Emilio [1 ]
Escalera, Sergio [1 ,2 ]
机构
[1] Univ Barcelona, Barcelona, Spain
[2] Comp Vis Ctr, Barcelona, Spain
来源
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) | 2021年
关键词
D O I
10.1109/ICCV48922.2021.00542
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We present a novel solution to the garment animation problem through deep learning. Our contribution allows animating any template outfit with arbitrary topology and geometric complexity. Recent works develop models for garment edition, resizing and animation at the same time by leveraging the support body model (encoding garments as body homotopies). This leads to complex engineering solutions that suffer from scalability, applicability and compatibility. By limiting our scope to garment animation only, we are able to propose a simple model that can animate any outfit, independently of its topology, vertex order or connectivity. Our proposed architecture maps outfits to animated 3D models into the standard format for 3D animation (blend weights and blend shapes matrices), automatically providing of compatibility with any graphics engine. We also propose a methodology to complement supervised learning with an unsupervised physically based learning that implicitly solves collisions and enhances cloth quality.
引用
收藏
页码:5451 / 5460
页数:10
相关论文
共 34 条
  • [11] Geometric Skinning with Approximate Dual Quaternion Blending
    Kavan, Ladislav
    Collins, Steven
    Zara, Jiri
    O'Sullivan, Carol
    [J]. ACM TRANSACTIONS ON GRAPHICS, 2008, 27 (04):
  • [12] Kavan Ladislav, 2005, P 2005 S INT 3D GRAP, P9, DOI DOI 10.1145/1053427.1053429
  • [13] DeepWrinkles: Accurate and Realistic Clothing Modeling
    Laehner, Zorah
    Cremers, Daniel
    Tung, Tony
    [J]. COMPUTER VISION - ECCV 2018, PT IV, 2018, 11208 : 698 - 715
  • [14] Smooth Skinning Decomposition with Rigid Bones
    Le, Binh Huy
    Deng, Zhigang
    [J]. ACM TRANSACTIONS ON GRAPHICS, 2012, 31 (06):
  • [15] Lewis JP, 2000, COMP GRAPH, P165, DOI 10.1145/344779.344862
  • [16] LIU T, 2017, ACM T GRAPHIC, V36, P1, DOI DOI 10.1145/3072959.2990496
  • [17] SMPL: A Skinned Multi-Person Linear Model
    Loper, Matthew
    Mahmood, Naureen
    Romero, Javier
    Pons-Moll, Gerard
    Black, Michael J.
    [J]. ACM TRANSACTIONS ON GRAPHICS, 2015, 34 (06):
  • [18] SMPLR: Deep learning based SMPL reverse for 3D human pose and shape recovery
    Madadi, Meysam
    Bertiche, Hugo
    Escalera, Sergio
    [J]. PATTERN RECOGNITION, 2020, 106
  • [19] Magnenat-Thalmann, 1988, Proceedings of Graphics Interface '88, P26
  • [20] Neural Body Fitting: Unifying Deep Learning and Model Based Human Pose and Shape Estimation
    Omran, Mohamed
    Lassner, Christoph
    Pons-Moll, Gerard
    Gehler, Peter V.
    Schiele, Bernt
    [J]. 2018 INTERNATIONAL CONFERENCE ON 3D VISION (3DV), 2018, : 484 - 494