Snapture - A Novel Neural Architecture for Combined Static and Dynamic Hand Gesture Recognition:TERAIS Data

This repository lists the datasets used for developing the experiments of the paper titled: Snapture - A Novel Neural Architecture for Combined Static and Dynamic Hand Gesture Recognition (Open access) by Hassan Ali, Doreen Jirak and Stefan Wermter. The study was conducted in the Knowledge Technology (WTM) group at the University of Hamburg.

GRIT Robot Commands Dataset

This dataset is was recorded at the Knowledge Technology (WTM) group at the University of Hamburg and can be requested here. The dataset is public and was not collected as part of this study.

Montalbano Co-Speech Dataset

This dataset was recorded as part of the ChaLearn Looking at People Challenge and can be downloaded from here. The dataset is public and was not collected as part of this study. The attached montalbano_segments.csv file can be used to create gesture segmentations of the test subset of this dataset.

Acknow ledgements

This work was partially supported by the DFG under project CML (TRR 169) and BMWK under project KI-SIGS and EU under project TERAIS.

Citation

To cite our paper, you can copy the following into your .bib file

@Article{Ali2023,
author={Ali, Hassan and Jirak, Doreen and Wermter, Stefan},
title={Snapture---a Novel Neural Architecture for Combined Static and Dynamic Hand Gesture Recognition},
journal={Cognitive Computation},
year={2023},
month={Jul},
day={17},
issn={1866-9964},
doi={10.1007/s12559-023-10174-z},
url={https://doi.org/10.1007/s12559-023-10174-z}
}