Deep learning on Lie groups for skeleton-based action recognition
Deep learning on Lie groups for skeleton-based action recognition
In recent years, skeleton-based action recognition has become a popular 3D classification problem. State-of-the-art methods typically first represent each motion sequence as a high-dimensional trajectory on a Lie group with an additional dynamic time warping, and then shallowly learn favorable Lie group features. In this paper we incorporate the Lie group structure into a deep network architecture to learn more appropriate Lie group features for 3D action recognition. Within the network structure, we design rotation mapping layers to transform the input Lie group features into desirable ones, which are aligned better in the temporal domain. To reduce the high feature dimensionality, the architecture is equipped with rotation pooling layers for the elements on the Lie group. Furthermore, we propose a logarithm mapping layer to map the resulting manifold data into a tangent space that facilitates the application of regular output layers for the final classification. Evaluations of the proposed network for standard 3D human action recognition datasets clearly demonstrate its superiority over existing shallow Lie group feature learning methods as well as most conventional deep learning methods.
1243-1252
Huang, Zhiwu
84f477cd-9097-44dd-a33e-ff71f253d36b
Wan, Chengde
d9762363-939d-4eb0-9c74-782e870d5d0d
Probst, Thomas
e6f3208e-af94-42d1-8990-39d7239c8531
Van Gool, Luc
fee5dbb5-1da6-46d7-85d8-628046091781
9 November 2017
Huang, Zhiwu
84f477cd-9097-44dd-a33e-ff71f253d36b
Wan, Chengde
d9762363-939d-4eb0-9c74-782e870d5d0d
Probst, Thomas
e6f3208e-af94-42d1-8990-39d7239c8531
Van Gool, Luc
fee5dbb5-1da6-46d7-85d8-628046091781
Huang, Zhiwu, Wan, Chengde, Probst, Thomas and Van Gool, Luc
(2017)
Deep learning on Lie groups for skeleton-based action recognition.
In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
IEEE.
.
(doi:10.1109/CVPR.2017.137).
Record type:
Conference or Workshop Item
(Paper)
Abstract
In recent years, skeleton-based action recognition has become a popular 3D classification problem. State-of-the-art methods typically first represent each motion sequence as a high-dimensional trajectory on a Lie group with an additional dynamic time warping, and then shallowly learn favorable Lie group features. In this paper we incorporate the Lie group structure into a deep network architecture to learn more appropriate Lie group features for 3D action recognition. Within the network structure, we design rotation mapping layers to transform the input Lie group features into desirable ones, which are aligned better in the temporal domain. To reduce the high feature dimensionality, the architecture is equipped with rotation pooling layers for the elements on the Lie group. Furthermore, we propose a logarithm mapping layer to map the resulting manifold data into a tangent space that facilitates the application of regular output layers for the final classification. Evaluations of the proposed network for standard 3D human action recognition datasets clearly demonstrate its superiority over existing shallow Lie group feature learning methods as well as most conventional deep learning methods.
This record has no associated files available for download.
More information
Published date: 9 November 2017
Identifiers
Local EPrints ID: 501235
URI: http://eprints.soton.ac.uk/id/eprint/501235
PURE UUID: a16990c5-ec56-4fbd-b240-f49a8854bbb3
Catalogue record
Date deposited: 27 May 2025 18:04
Last modified: 28 May 2025 02:12
Export record
Altmetrics
Contributors
Author:
Zhiwu Huang
Author:
Chengde Wan
Author:
Thomas Probst
Author:
Luc Van Gool
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics