TY - GEN
T1 - PointCNN-Hand
T2 - 2021 International Conference on System Science and Engineering, ICSSE 2021
AU - Chen, Jia Hong
AU - Hsu, Chen Chien
N1 - Funding Information:
V. ACKOWLEDGMENT This work was financially supported by the “Chinese Language and Technology Center” of National Taiwan Normal University (NTNU) from The Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan, and Ministry of Science and Technology, Taiwan, under Grants no. MOST 110-2634-F-003-006 and MOST 110-2634-F-003-007 through Pervasive Artificial Intelligence Research (PAIR) Labs. We are grateful to the National Center for High-performance Computing for computer time and facilities to conduct this research.
Publisher Copyright:
© 2021 IEEE.
PY - 2021/8/26
Y1 - 2021/8/26
N2 - This paper provides a novel method called "PointCNN-Hand"for 3D hand joints estimation based on PointCNN. To use the depth image effectively, we transfer the hand depth image into the 3D hand cloud point and implement end-to-end training by PointCNN-Hand for hand joint estimation. We then perform error analysis on MSRA, NYU, and ICVL datasets to compare with the state-of-the-art methods. The experiments show that the proposed method has desired results, and the model parameters are relatively smaller than those of other methods. To be specific, the parameters of the proposed PointCNN-Hand network are reduced to only 3 Mega Byte (MB) with Floating Point Operations (FLOPs) less than 232.05M.
AB - This paper provides a novel method called "PointCNN-Hand"for 3D hand joints estimation based on PointCNN. To use the depth image effectively, we transfer the hand depth image into the 3D hand cloud point and implement end-to-end training by PointCNN-Hand for hand joint estimation. We then perform error analysis on MSRA, NYU, and ICVL datasets to compare with the state-of-the-art methods. The experiments show that the proposed method has desired results, and the model parameters are relatively smaller than those of other methods. To be specific, the parameters of the proposed PointCNN-Hand network are reduced to only 3 Mega Byte (MB) with Floating Point Operations (FLOPs) less than 232.05M.
KW - 3D hand pose estimation
KW - Convolutional Neural Network
KW - Hand articulation
UR - http://www.scopus.com/inward/record.url?scp=85116229372&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85116229372&partnerID=8YFLogxK
U2 - 10.1109/ICSSE52999.2021.9538459
DO - 10.1109/ICSSE52999.2021.9538459
M3 - Conference contribution
AN - SCOPUS:85116229372
T3 - Proceedings of 2021 International Conference on System Science and Engineering, ICSSE 2021
SP - 458
EP - 463
BT - Proceedings of 2021 International Conference on System Science and Engineering, ICSSE 2021
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 26 August 2021 through 28 August 2021
ER -