Name
Title | ||
|---|---|---|
Generating Hand Posture and Motion Dataset for Hand Pose Estimation in Egocentric View |
Abstract | ||
|---|---|---|
Hand interaction is one of the main input modalities for augmented reality glasses. Vision-based approaches using deep learning have been applied to hand tracking and shown good results. To train a deep neural network, a large dataset of hand information is required. However, obtaining real hand data is painful due to a large number of annotations and lack of diversities such as skins, lighting conditions, and backgrounds. In this paper, we propose a method to generate a synthetic hand dataset that includes diverse human and environmental parameters. By applying constraints of a human hand, we can get realistic hand poses for hand dataset. We also generate dynamic hand animations which can be used for hand gesture recognition. |
Year | DOI | Venue |
|---|---|---|
2022 | 10.1007/978-3-031-05939-1_22 | Virtual, Augmented and Mixed Reality: Design and Development |
Keywords | DocType | Volume |
Hand pose estimation, Hand dataset, Synthetic dataset | Conference | 13317 |
ISSN | Citations | PageRank |
0302-9743 | 0 | 0.34 |
References | Authors | |
0 | 9 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Park Hwangpil | 1 | 0 | 0.34 |
| Kim Deokho | 2 | 0 | 0.34 |
| Yim Sunghoon | 3 | 0 | 0.34 |
| Kwon Taehyuk | 4 | 0 | 0.34 |
| Jeong Jiwon | 5 | 0 | 0.34 |
| Lee Wonwoo | 6 | 0 | 0.34 |
| Lee Jaewoong | 7 | 0 | 0.34 |
| Yoo Byeongwook | 8 | 0 | 0.34 |
| Lee Gunill | 9 | 0 | 0.34 |