matlab code for 3D skeleton action recognition
by DING Chong-Yang and LIQ Kai(http://web.xidian.edu.cn/kailiu/).
Raviteja Vemulapalli, Felipe Arrate, and Rama Chellappa, "Human Action Recognition by Representing 3D Human Skeletons as Points in a Lie Group", CVPR, 2014.
please cite it if you are going to use it.
The code is implemented by Raviteja Vemulapalli and developed by us to mainly comparing the performance of some different feature-extract methods, such as Joint locations, Joint angles, Lie group and STWP(proposed by us, and turned out to be a better approach compared with others).
This code has been implemented in Matlab R2017a and tested in both Linux (ubuntu) and Windows 7.
Cross-subject - half of the subjects used for training and the remaining half used for testing. Results are averaged over 10 different training and test subject combinations.
We provide pre-computed skeleton sequences for all the datasets supported:
The matlab file "run.m" runs the experiments for UTKinect-Action, Florence3D-Action and MSRAction3D datasets using 5 different skeletal representations: 'JP', 'RJP', 'JA', 'Lie Group' and 'STWP'.
The file "skeletal_action_classification.m" contains the code for entire pipeline: Step 1: Skeletal representation ('JP' or 'RJP' or 'JA' or 'Lie Group' or 'STWP') Step 2: Temporal modeling (DTW and Fourier Temporal Pyramid) Step 3: Classification: One-vs-All linear SVM (implemented as kernel SVM with linear kernel)
On the MSRAction3D dataset, which we have downloaded into this file, the recognition rates of different skeletal representations: 'JP', 'RJP', 'JA', 'Lie Group' and 'STWP' are:
88.75%, 88.87%, 75.39%, 89.55%, 92.79%
For any query please contact us for more information (kailiu@mail.xidian.edu.cn).