Spatio-Temporal Image Encoding

This repository contains the implementation of the Spatio-Temporal Image Encoding (STIE), in order to perform Online Human Activity Recognition using 3D skeletons. This method encodes a sequence of 3D skeletons into an image, while preserving both spatial and temporal dependencies.

Our paper can be found at:

Human Activity Recognition: A Spatio-temporal Image Encoding of 3D Skeleton Data for Online Action Detection

If you use or build on our work, please consider citing us:

@conference{visapp22,  
author={Nassim Mokhtari. and Alexis Nédélec. and Pierre {De Loor}.},  
title={Human Activity Recognition: A Spatio-temporal Image Encoding of 3D Skeleton Data for Online Action Detection},  
booktitle={Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 5: VISAPP},  
year={2022},  
pages={448-455},  
publisher={SciTePress},  
organization={INSTICC},  
doi={10.5220/0010835800003124},  
isbn={978-989-758-555-5},  
issn={2184-4321},  
}

Dataset

Before running our code, please unzip the archive data.zip provided in this repo. This archive contains skeleton data and sequence labels from the Online Action Detection dataset.

note: If you are using your own dataset, please consider adjusting the load_data_file() function.

Usage

You can start the encoding using the default parameters by running the STIE.py from the command line :

python ./STIE.py

Several parameters can be used to adapt the encoding according to your needs. You can find more details about these parameters using :

python ./STIE.py --help

Encoded Sequence

Real time detection

Results

Using pre-trained VGG16