Time-sensitive topic models for action recognition in videos

Romain Tavenard, Rémi Emonet, Jean-Marc Odobez

Abstract

In this paper, we postulate that temporal information is important for action recognition in videos. Keeping temporal information, videos are represented as word×time documents. We propose to use time-sensitive probabilistic topic models and we extend them for the context of supervised learning. Our time-sensitive approach is compared to both PLSA and Bag-of-Words. Our approach is shown to both capture semantics from data and yield classification performance comparable to other methods, outperforming them when the amount of training data is low.
BibTeX
@inproceedings{tavenard2013time,
  author = {Romain Tavenard and Rémi Emonet and Jean-Marc Odobez},
  title = {{Time-sensitive topic models for action recognition in videos}},
  booktitle = {ICIP},
  year = {2013},
  doi = {10.1109/icip.2013.6738615},
}

Loading related content...

Figure