Hear Me out: Fusional Approaches for Audio Augmented Temporal Action Localization

Anurag Bagchi, Jazib Mahmood, Dolton Fernandes, Ravi Kiran Sarvadevabhatla

2022

Abstract

State of the art architectures for untrimmed video Temporal Action Localization (TAL) have only considered RGB and Flow modalities, leaving the information-rich audio modality unexploited. Audio fusion has been explored for the related but an arguably easier problem of trimmed (clip-level) action recognition. However, TAL poses a unique set of challenges. In this paper, we propose simple but effective fusion-based approaches for TAL. To the best of our knowledge, our work is the first to jointly consider audio and video modalities for supervised TAL. We experimentally show that our schemes consistently improve performance for the state-of-the-art video-only TAL approaches. Specifically, they help achieve a new state-of-the-art performance on large-scale benchmark datasets - ActivityNet-1.3 (54.34 mAP@0.5) and THUMOS14 (57.18 mAP@0.5). Our experiments include ablations involving multiple fusion schemes, modality combinations, and TAL architectures. Our code, models, and associated data are available at https://github.com/skelemoa/tal-hmo.

Paper Citation

in Harvard Style

Bagchi A., Mahmood J., Fernandes D. and Sarvadevabhatla R. (2022). Hear Me out: Fusional Approaches for Audio Augmented Temporal Action Localization. In Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 5: VISAPP; ISBN 978-989-758-555-5, SciTePress, pages 144-154. DOI: 10.5220/0010832700003124

in Bibtex Style

@conference{visapp22,
author={Anurag Bagchi and Jazib Mahmood and Dolton Fernandes and Ravi Kiran Sarvadevabhatla},
title={Hear Me out: Fusional Approaches for Audio Augmented Temporal Action Localization},
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={144-154},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010832700003124},
isbn={978-989-758-555-5},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 5: VISAPP
TI - Hear Me out: Fusional Approaches for Audio Augmented Temporal Action Localization
SN - 978-989-758-555-5
AU - Bagchi A.
AU - Mahmood J.
AU - Fernandes D.
AU - Sarvadevabhatla R.
PY - 2022
SP - 144
EP - 154
DO - 10.5220/0010832700003124
PB - SciTePress