ActionHOG is an open source C++ implementation of an efficient local spatio-temporal detector and descriptor for human activity analysis.

To compile the code, you need to have the OpenCV library (tested with OpenCV-2.3.1).

./ActionHOG -i vidFile -o featFile [options]

The configurations and parameters can be set in the options.

If you find any bug or develop some extensions, please feel free to drop me a line.

Please cite our paper if you use the code:

X. Yang, C. Yi, L. Cao, and Y. Tian. MediaCCNY at TRECVID 2012: Surveillance Event Detection. NIST TRECVID Workshop, 2012. [PDF]

Y. Tian, L. Cao, Z. Liu, and Z. Zhang. Hierarchical Filtered Motion for Action Recognition in Crowded Videos. IEEE Transactions on Systems, Man, and Cybernetics (Part C) 42(3), 313-323, 2012. [PDF]

Copyright (C) 2013 Xiaodong Yang

Distibution code version 1.0 - 05/10/2014. The code is for research purpose only.