Human activity recognition from video sequences

Activity recognition by deep learning

Deep attention models for activity recognition. Work of Fabien Bardel.

Papers:

We present deep models for spatio-temporal data designed for activity recognition in videos. An emphasis is put on unsupervised learning of invariant models.

Papers:

Learning binary motion patterns with deep belief networks.

Papers:

The LIRIS HARL dataset

The LIRIS human activities dataset and the ICPR HARL 2012 competition.

Papers:

Activity recognition by graph matching

Papers:

Activity recognition by BoW models

Papers:

Contact

Christian Wolf
LIRIS UMR CNRS 5205
INSA-Lyon / Université de Lyon