Domain Generalization for Activity Recognition: Learn from Visible, Infer with Thermal

Yannick Zoetgnande, Jean Dillenseger

2022

Abstract

We proposed a solution based on I3D and optical flow to learn common characteristics between thermal and visible videos. For this purpose we proposed a new database to evaluate our solution. The new model comprises an optical flow extractor; a feature extractor based on I3D, a domain classifier, and an activity recognition classifier. We learn invariant characteristics computed from the optical flow. We have simulated several source domains, and we have shown that it is possible to obtain excellent results on a modality that was not used during the training. Such techniques can be used when there is only one source and one target domain.

Paper Citation

in Harvard Style

Zoetgnande Y. and Dillenseger J. (2022). Domain Generalization for Activity Recognition: Learn from Visible, Infer with Thermal. In Proceedings of the 11th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-549-4, pages 722-729. DOI: 10.5220/0010906300003122

in Bibtex Style

@conference{icpram22,
author={Yannick Zoetgnande and Jean Dillenseger},
title={Domain Generalization for Activity Recognition: Learn from Visible, Infer with Thermal},
booktitle={Proceedings of the 11th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2022},
pages={722-729},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010906300003122},
isbn={978-989-758-549-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 11th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - Domain Generalization for Activity Recognition: Learn from Visible, Infer with Thermal
SN - 978-989-758-549-4
AU - Zoetgnande Y.
AU - Dillenseger J.
PY - 2022
SP - 722
EP - 729
DO - 10.5220/0010906300003122