International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2020
Hand-3D-Studio:A New Multi-View System For 3D Hand Reconstruction
Zhengyi Zhao,
  
Tianyao Wang,
  
Siyu Xia,
  
Yangang Wang
Southeast University
Hardware System

  • Resolution: 15 DSLR cameras acquired high quality multi-view 4K resolution color images in a circular manner.
  • Ergonomic design: the height and angle of the customized holder can be adjusted depending on the height and shape human.
  • Synchronization: The synchronization delay is less than 10 milliseconds, far less than the USB software control scheme.
  • Lighting: A total length of 50 meters LED strips are mounted on the outsideofthe”yurt”to provide a shadowless illumination around the hand.
Multi-view 4K Dataset

We have collected multiview hand color images with 10 persons of different genders and skin colors. The gestures we collected are very representative and common in daily life. For each person, we collected 50 one-handed gestures and 27 hand-object interaction gestures, both for the left and right hand. The estures can be divided into 3 categories: finger-movements, common gestures and hand-object interaction gestures.
Multi-view 4K Dataset
Attributes Value
num.subjects 10
num.objects 27
num.cameras 15
num.frames 22K/-
resolution 4K
3D joints annotation
3D shape annotation
Download

The dataset has a total size of 102GB. For your convenience, the dataset has been cut into 10 parts to
Reference

Zhengyi Zhao, Tianyao Wang, Siyu Xia and Yangang Wang. "Hand-3D-Studio:A New Multi-View System For 3D Hand Reconstruction". International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2020.
Acknowledgments: This work was supported in part by the National Natural Science Foundation of China (No. 61806054), in part by the Natural Science Foundation of Jiangsu Province (No. BK20180355), in part by the Shenzhen Science and Technology Innovation Committee (STIC) (JCYJ20180306174459972) and “Zhishan Young Scholar” Program of Southeast University.