Domain-Specific Priors and Meta Learning for Few-Shot First-Person Action Recognition
IEEE Transactions on Pattern Analysis and Machine Intelligence · 01.06.2023 · 45 (6) · 6659-6673
DOI: 10.1109/tpami.2021.3058606 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, erschienen am 01.06.2023. Verfasst von Huseyin Coskun und 6 weiteren Autor:innen. Hauptthemen laut MEDLINE: Algorithms, Learning. Die Arbeit wurde bisher 25-mal zitiert (Crossref) und nennt 87 Literaturangaben. Gefördert von: Microsoft.
Abstract (Originalsprache)
The lack of large-scale real datasets with annotations makes transfer learning a necessity for video activity understanding. We aim to develop an effective method for few-shot transfer learning for first-person action classification. We leverage independently trained local visual cues to learn representations that can be transferred from a source domain, which provides primitive action labels, to a different target domain - using only a handful of examples. Visual cues we employ include object-object interactions, hand grasps and motion within regions that are a function of hand locations. We employ a framework based on meta-learning to extract the distinctive and domain invariant components of the deployed visual cues. This enables transfer of action classification models across public datasets captured with diverse scene and action configurations. We present comparative results of our transfer learning methodology and report superior results over state-of-the-art action classification approaches for both inter-class and inter-dataset transfer.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Förderung
- Microsoft
Zitieren
APA
Coskun, H., Zia, M. Z., Tekin, B., Bogo, F., Navab, N., Tombari, F., & Sawhney, H. S. (2023). Domain-Specific Priors and Meta Learning for Few-Shot First-Person Action Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(6), 6659-6673. https://doi.org/10.1109/tpami.2021.3058606
Vancouver
Coskun H, Zia MZ, Tekin B, Bogo F, Navab N, Tombari F, et al. Domain-Specific Priors and Meta Learning for Few-Shot First-Person Action Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2023;45(6):6659-6673. doi:10.1109/tpami.2021.3058606
BibTeX
@article{Coskun2023,
author = {Coskun, Huseyin and Zia, M Zeeshan and Tekin, Bugra and Bogo, Federica and Navab, Nassir and Tombari, Federico and Sawhney, Harpreet S},
title = {Domain-Specific Priors and Meta Learning for Few-Shot First-Person Action Recognition},
journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
year = {2023},
volume = {45},
number = {6},
pages = {6659--6673},
doi = {10.1109/tpami.2021.3058606}
}
EUVIMED