GarTemFormer: Temporal transformer-based for optimizing virtual garment animation

Virtual garment animation and deformation constitute a pivotal research direction in computer graphics, finding extensive applications in domains such as computer games, animation, and film. Traditional physics-based methods can simulate the physical characteristics of garments, such as elasticity a...

Full description

Saved in:
Bibliographic Details
Main Authors: Jiazhe Miao, Tao Peng, Fei Fang, Xinrong Hu, Li Li
Format: Article
Language:English
Published: Elsevier 2024-12-01
Series:Graphical Models
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S1524070324000237
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846138599036157952
author Jiazhe Miao
Tao Peng
Fei Fang
Xinrong Hu
Li Li
author_facet Jiazhe Miao
Tao Peng
Fei Fang
Xinrong Hu
Li Li
author_sort Jiazhe Miao
collection DOAJ
description Virtual garment animation and deformation constitute a pivotal research direction in computer graphics, finding extensive applications in domains such as computer games, animation, and film. Traditional physics-based methods can simulate the physical characteristics of garments, such as elasticity and gravity, to generate realistic deformation effects. However, the computational complexity of such methods hinders real-time animation generation. Data-driven approaches, on the other hand, learn from existing garment deformation data, enabling rapid animation generation. Nevertheless, animations produced using this approach often lack realism, struggling to capture subtle variations in garment behavior. We proposes an approach that balances realism and speed, by considering both spatial and temporal dimensions, we leverage real-world videos to capture human motion and garment deformation, thereby producing more realistic animation effects. We address the complexity of spatiotemporal attention by aligning input features and calculating spatiotemporal attention at each spatial position in a batch-wise manner. For garment deformation, garment segmentation techniques are employed to extract garment templates from videos. Subsequently, leveraging our designed Transformer-based temporal framework, we capture the correlation between garment deformation and human body shape features, as well as frame-level dependencies. Furthermore, we utilize a feature fusion strategy to merge shape and motion features, addressing penetration issues between clothing and the human body through post-processing, thus generating collision-free garment deformation sequences. Qualitative and quantitative experiments demonstrate the superiority of our approach over existing methods, efficiently producing temporally coherent and realistic dynamic garment deformations.
format Article
id doaj-art-4e2041350ff646049f6d9355024fb6fd
institution Kabale University
issn 1524-0703
language English
publishDate 2024-12-01
publisher Elsevier
record_format Article
series Graphical Models
spelling doaj-art-4e2041350ff646049f6d9355024fb6fd2024-12-07T08:25:15ZengElsevierGraphical Models1524-07032024-12-01136101235GarTemFormer: Temporal transformer-based for optimizing virtual garment animationJiazhe Miao0Tao Peng1Fei Fang2Xinrong Hu3Li Li4School of Computer and Artificial Intelligence, Wuhan Textile University, No. 1 Sunshine Avenue, Jiangxia District, Wuhan, 430200, Hubei Province, China; Engineering Research Center of Hubei Province for Clothing Information, Wuhan, 430200, Hubei Province, ChinaSchool of Computer and Artificial Intelligence, Wuhan Textile University, No. 1 Sunshine Avenue, Jiangxia District, Wuhan, 430200, Hubei Province, China; Engineering Research Center of Hubei Province for Clothing Information, Wuhan, 430200, Hubei Province, China; China National Textile and Apparel Council Key Laboratory of Intelligent Perception and Computing, Wuhan, 430200, Hubei Province, China; Corresponding author at: School of Computer and Artificial Intelligence, Wuhan Textile University, No. 1 Sunshine Avenue, Jiangxia District, Wuhan, 430200, Hubei Province, China.School of Computer and Artificial Intelligence, Wuhan Textile University, No. 1 Sunshine Avenue, Jiangxia District, Wuhan, 430200, Hubei Province, China; Engineering Research Center of Hubei Province for Clothing Information, Wuhan, 430200, Hubei Province, ChinaSchool of Computer and Artificial Intelligence, Wuhan Textile University, No. 1 Sunshine Avenue, Jiangxia District, Wuhan, 430200, Hubei Province, China; China National Textile and Apparel Council Key Laboratory of Intelligent Perception and Computing, Wuhan, 430200, Hubei Province, ChinaSchool of Computer and Artificial Intelligence, Wuhan Textile University, No. 1 Sunshine Avenue, Jiangxia District, Wuhan, 430200, Hubei Province, China; Engineering Research Center of Hubei Province for Clothing Information, Wuhan, 430200, Hubei Province, ChinaVirtual garment animation and deformation constitute a pivotal research direction in computer graphics, finding extensive applications in domains such as computer games, animation, and film. Traditional physics-based methods can simulate the physical characteristics of garments, such as elasticity and gravity, to generate realistic deformation effects. However, the computational complexity of such methods hinders real-time animation generation. Data-driven approaches, on the other hand, learn from existing garment deformation data, enabling rapid animation generation. Nevertheless, animations produced using this approach often lack realism, struggling to capture subtle variations in garment behavior. We proposes an approach that balances realism and speed, by considering both spatial and temporal dimensions, we leverage real-world videos to capture human motion and garment deformation, thereby producing more realistic animation effects. We address the complexity of spatiotemporal attention by aligning input features and calculating spatiotemporal attention at each spatial position in a batch-wise manner. For garment deformation, garment segmentation techniques are employed to extract garment templates from videos. Subsequently, leveraging our designed Transformer-based temporal framework, we capture the correlation between garment deformation and human body shape features, as well as frame-level dependencies. Furthermore, we utilize a feature fusion strategy to merge shape and motion features, addressing penetration issues between clothing and the human body through post-processing, thus generating collision-free garment deformation sequences. Qualitative and quantitative experiments demonstrate the superiority of our approach over existing methods, efficiently producing temporally coherent and realistic dynamic garment deformations.http://www.sciencedirect.com/science/article/pii/S1524070324000237SimulationPhysical constraintsVirtual try-onMotion sequences
spellingShingle Jiazhe Miao
Tao Peng
Fei Fang
Xinrong Hu
Li Li
GarTemFormer: Temporal transformer-based for optimizing virtual garment animation
Graphical Models
Simulation
Physical constraints
Virtual try-on
Motion sequences
title GarTemFormer: Temporal transformer-based for optimizing virtual garment animation
title_full GarTemFormer: Temporal transformer-based for optimizing virtual garment animation
title_fullStr GarTemFormer: Temporal transformer-based for optimizing virtual garment animation
title_full_unstemmed GarTemFormer: Temporal transformer-based for optimizing virtual garment animation
title_short GarTemFormer: Temporal transformer-based for optimizing virtual garment animation
title_sort gartemformer temporal transformer based for optimizing virtual garment animation
topic Simulation
Physical constraints
Virtual try-on
Motion sequences
url http://www.sciencedirect.com/science/article/pii/S1524070324000237
work_keys_str_mv AT jiazhemiao gartemformertemporaltransformerbasedforoptimizingvirtualgarmentanimation
AT taopeng gartemformertemporaltransformerbasedforoptimizingvirtualgarmentanimation
AT feifang gartemformertemporaltransformerbasedforoptimizingvirtualgarmentanimation
AT xinronghu gartemformertemporaltransformerbasedforoptimizingvirtualgarmentanimation
AT lili gartemformertemporaltransformerbasedforoptimizingvirtualgarmentanimation