Institute Homepage
Institute Homepage DE Sign In

Back

Perceiving Systems Conference Paper 2022

DART: Articulated Hand Model with Diverse Accessories and Rich Textures

Thumb ticker sm 84b79ecc0b10a9579306eaa7c8037a
Perceiving Systems
Yuliang Xiu
  • Guest Scientist
Teaser

Hand, the bearer of human productivity and intelligence, is receiving much attention due to the recent fever of 3D digital avatars. Among different hand morphable models, MANO has been widely used in various vision & graphics tasks. However, MANO disregards textures and accessories, which largely limits its power to synthesize photorealistic & lifestyle hand data. In this paper, we extend MANO with more Diverse Accessories and Rich Textures, namely DART. DART is comprised of 325 exquisite hand-crafted texture maps which vary in appearance and cover different kinds of blemishes, make-ups, and accessories. We also provide the Unity GUI which allows people to render hands with user-specific settings, e.g. pose, camera, background, lighting, and DART textures. In this way, we generate large-scale (800K), diverse, and high-fidelity hand images, paired with perfect-aligned 3D labels, called DARTset. Experiments demonstrate its superiority in generalization and diversity. As a great complement to existing datasets, DARTset could boost hand pose estimation & surface reconstruction tasks. DART and Unity software will be publicly available for research purposes.

Author(s): Gao, Daiheng* and Xiu, Yuliang* and Li, Kailin* and Yang, Lixin* and Wang, Feng and Zhang, Peng and Zhang, Bang and Lu, Cewu and Tan, Ping
Book Title: Thirty-sixth Conference on Neural Information Processing Systems (NeurIPS)
Year: 2022
Month: November
Bibtex Type: Conference Paper (conference)
Event Name: NeurIPS 2022-Datasets and Benchmarks Track
Event Place: New Orleans, Louisiana
State: Published
Electronic Archiving: grant_archive
Links:

BibTex 

@conference{dart2022,
  title = {{DART}: {A}rticulated {H}and {M}odel with {D}iverse {A}ccessories and {R}ich {T}extures},
  booktitle = {Thirty-sixth Conference on Neural Information Processing Systems (NeurIPS)},
  abstract = {Hand, the bearer of human productivity and intelligence, is receiving much attention due to the recent fever of 3D digital avatars. Among different hand morphable models, MANO has been widely used in various vision & graphics tasks. However, MANO disregards textures and accessories, which largely limits its power to synthesize photorealistic & lifestyle hand data. In this paper, we extend MANO with more Diverse Accessories and Rich Textures, namely DART. DART is comprised of 325 exquisite hand-crafted texture maps which vary in appearance and cover different kinds of blemishes, make-ups, and accessories. We also provide the Unity GUI which allows people to render hands with user-specific settings, e.g. pose, camera, background, lighting, and DART textures. In this way, we generate large-scale (800K), diverse, and high-fidelity hand images, paired with perfect-aligned 3D labels, called DARTset. Experiments demonstrate its superiority in generalization and diversity. As a great complement to existing datasets, DARTset could boost hand pose estimation & surface reconstruction tasks. DART and Unity software will be publicly available for research purposes.},
  month = nov,
  year = {2022},
  slug = {dart2022},
  author = {Gao, Daiheng* and Xiu, Yuliang* and Li, Kailin* and Yang, Lixin* and Wang, Feng and Zhang, Peng and Zhang, Bang and Lu, Cewu and Tan, Ping},
  month_numeric = {11}
}