Predicting Loose-Fitting Garment Deformations Using Bone-Driven Motion Networks

Pan, Xiaoyu; Mai, Jiaming; Jiang, Xinwei; Tang, Dongxue; Li, Jingxiang; Shao, Tianjia; Zhou, Kun; Jin, Xiaogang; Manocha, Dinesh

doi:10.1145/3528233.3530709

Cited by 20 publications

(26 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[WWW22] propose a GPU‐based multilevel additive Schwarz preconditioner to simulate cloth with a high resolution, 50K to 500K vertices, in real‐time. On the other hand, alternative methods have also been developed to generate the dynamic details of cloth animation via adaptive techniques [LYO*10, MC10, NSO12], data‐driven approaches [dASTH10, GRH*12, WHRO10, KGBS11, ZBO13] and deep learning‐based methods [CYJ*18, CZY21, GCS*19, LCT18, ZWCM20, BME20, GCP*20,SOC22,PMJ*22,TB23,ZWCM21,ZCM22], etc.…”

Section: Related Workmentioning

confidence: 99%

“…These approaches add wrinkles on normal maps rather than geometry, and thus their effectiveness is restricted to adding fine‐scale visual details, not large‐scale dynamics. Based on the skinning representation, there is a tremendous amuount of research focusing on body‐ or skeleton‐guided garment generation with neural networks, which aims to generalize to multiple body shapes [GCS*19, SOC19, GCP*20], loose‐fitting garments [PMJ*22], semi‐supervised or unsupervised generation [ZCM22, BME20, SOC22]. In addition, other works are devoted to generalizing neural networks to various cloth styles [PLPM20] or cloth materials [WSFM19].…”

Section: Related Workmentioning

confidence: 99%

“…Recently, data‐driven methods provide alternative solutions to fast and effective wrinkling behaviors for garments. Depending on body poses, some data‐driven methods [WHRO10,FYK10, dASTH10, SOC19, WSFM19, PMJ*22] are capable of generating tight or loose‐fitting cloth animations successfully.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Deep Deformation Detail Synthesis for Thin Shell Models

Chen

Gao

Yang

et al. 2023

Computer Graphics Forum

View full text Add to dashboard Cite

In physics‐based cloth animation, rich folds and detailed wrinkles are achieved at the cost of expensive computational resources and huge labor tuning. Data‐driven techniques make efforts to reduce the computation significantly by utilizing a preprocessed database. One type of methods relies on human poses to synthesize fitted garments, but these methods cannot be applied to general cloth animations. Another type of methods adds details to the coarse meshes obtained through simulation, which does not have such restrictions. However, existing works usually utilize coordinate‐based representations which cannot cope with large‐scale deformation, and requires dense vertex correspondences between coarse and fine meshes. Moreover, as such methods only add details, they require coarse meshes to be sufficiently close to fine meshes, which can be either impossible, or require unrealistic constraints to be applied when generating fine meshes. To address these challenges, we develop a temporally and spatially as‐consistent‐as‐possible deformation representation (named TS‐ACAP) and design a DeformTransformer network to learn the mapping from low‐resolution meshes to ones with fine details. This TS‐ACAP representation is designed to ensure both spatial and temporal consistency for sequential large‐scale deformations from cloth animations. With this TS‐ACAP representation, our DeformTransformer network first utilizes two mesh‐based encoders to extract the coarse and fine features using shared convolutional kernels, respectively. To transduct the coarse features to the fine ones, we leverage the spatial and temporal Transformer network that consists of vertex‐level and frame‐level attention mechanisms to ensure detail enhancement and temporal coherence of the prediction. Experimental results show that our method is able to produce reliable and realistic animations in various datasets at high frame rates with superior detail synthesis abilities compared to existing methods.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Deep Deformation Detail Synthesis for Thin Shell Models

Chen

Gao

Yang

et al. 2023

Computer Graphics Forum

View full text Add to dashboard Cite

show abstract

“…Physics-based ones [3,32,33,34,35,12] produce high-quality results but are computationally demanding, while data-driven approaches are faster, sometimes at the cost of realism. Most datadriven methods are template-based [13,14,15,17,19,18,20,16,18,13], with a triangulated mesh modeling a specific garment and a draping function trained specifically for it. As this is impractical for large garment collections, some recent works [22,21,36] use 3D point clouds to represent garments instead.…”

Section: Related Workmentioning

confidence: 99%

“…Hence, many recent techniques use neural networks to speed up the draping and to make it differentiable. The garments can be represented by 3D mesh templates [13,14,15,16,17,18,19,20], point clouds [21,22], UV maps [23,24,25,26,27], or implicit surfaces [28,29,30]. Draping can then be achieved by Linear Blend Skinning (LBS) from the shape and pose parameters of a body model, such as SMPL [31].…”

Section: Introductionmentioning

confidence: 99%

DIG: Draping Implicit Garment over the Human Body

Ren

Guillard

Remelli³

et al. 2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Many approaches to draping individual garments on human body models are realistic, fast, and yield outputs that are differentiable with respect to the body shape on which they are draped. However, none of them can handle multi-layered clothing, which is prevalent in everyday dress. In this paper, we introduce a parametric garment representation model that can. As in models used by clothing designers, each garment consists of individual 2D panels. Their 2D shape is defined by a Signed Distance Function and 3D shape by a 2D to 3D mapping. The 2D parameterization enables easy detection of potential collisions and the 3D parameterization handles complex shapes effectively. We show that this combination is faster and yields higher quality reconstructions than purely implicit surface representations, and makes the recovery of layered garments from images possible thanks to its differentiability. Furthermore, it supports rapid editing of garment shapes and texture by modifying individual 2D panels.Preprint. Under review.

show abstract