Deep Learning from Temporal Coherence in Video

Abstract

This work proposes a learning method for deep architectures that takes advantage of sequential data, in particular from the temporal coherence that naturally exists in unlabeled video recordings. That is, two successive frames are likely to contain the same object or objects. This coherence is used as a supervisory signal over the unlabeled data, and is used to improve the performance on a supervised task of interest. We demonstrate the effectiveness of this method on some pose invariant object and face recognition tasks.

Cite

Text

Mobahi et al. "Deep Learning from Temporal Coherence in Video." International Conference on Machine Learning, 2009. doi:10.1145/1553374.1553469

Markdown

[Mobahi et al. "Deep Learning from Temporal Coherence in Video." International Conference on Machine Learning, 2009.](https://mlanthology.org/icml/2009/mobahi2009icml-deep/) doi:10.1145/1553374.1553469

BibTeX

@inproceedings{mobahi2009icml-deep,
  title     = {{Deep Learning from Temporal Coherence in Video}},
  author    = {Mobahi, Hossein and Collobert, Ronan and Weston, Jason},
  booktitle = {International Conference on Machine Learning},
  year      = {2009},
  pages     = {737-744},
  doi       = {10.1145/1553374.1553469},
  url       = {https://mlanthology.org/icml/2009/mobahi2009icml-deep/}
}