EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Interpretability in Video-based Human Action Recognition: Saliency Maps and GradCAM in 3D Convolutional Neural Networks

Jorge Garcia-Torres Fernandez, Øyvind Meinich-Bache

Nordic Machine Intelligence · 2024

DOI: 10.5617/nmi.10587 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Nordic Machine Intelligence, 2024. Verfasst von Jorge Garcia-Torres Fernandez und 1 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Interpretability plays a vital role in understanding complex deep learning models by providing transparency and insights. It addresses the black-box nature of these models, aids in human-in-the-loop systems, enhances model development, and supports education. However, existing interpretability algorithms in computer vision primarily target images, leaving a gap for video applications. In this article, we emphasize the importance of interpretability in video-based Human Action Recognition (HAR). We extend existing 2D interpretability techniques to the 3D domain, specifically focusing on saliency maps and gradient-weighted class activation maps. The proposed interpretability system is then employed in the analysis of a well-known HAR dataset to better understand action recognition in videos.

Zitieren

APA

Fernandez, J. G., & Meinich-Bache, Ø. (2024). Interpretability in Video-based Human Action Recognition: Saliency Maps and GradCAM in 3D Convolutional Neural Networks. Nordic Machine Intelligence. https://doi.org/10.5617/nmi.10587

Vancouver

Fernandez JG, Meinich-Bache Ø. Interpretability in Video-based Human Action Recognition: Saliency Maps and GradCAM in 3D Convolutional Neural Networks. Nordic Machine Intelligence. 2024. doi:10.5617/nmi.10587

BibTeX
@article{Fernandez2024,
  author = {Fernandez, Jorge Garcia-Torres and Meinich-Bache, Øyvind},
  title = {Interpretability in Video-based Human Action Recognition: Saliency Maps and GradCAM in 3D Convolutional Neural Networks},
  journal = {Nordic Machine Intelligence},
  year = {2024},
  doi = {10.5617/nmi.10587}
}
Kostenlos starten →