Microsoft Research-INRIA joint center, Paris, France
and
LEAR team,
INRIA, Grenoble, France
|
655 Avenue de l'Europe 38330 Montbonnot France |
Phone: Office: E-mail: |
+33 4 7661 5441 H 215 firstname.lastname@inria.fr |
| We propose a new kernel specific to videos, which compares actions as time series of frames. The dynamics of actions are compared by using the Hilbert-Schmidt norm of the difference between their respective auto-correlations, in a RKHS induced by a kernel between frames. Our time series kernel (called DACO, for Difference of Auto-Correlation Operators) is shown to provide complementary information to bag-of-features. The results are published at the BMVC 2011 conference in the paper A time series kernel for action recognition A one page summary is also available here. |
|
| More details coming soon... |
| The goal of this work is to find when a short human action (like opening a door) is performed in large real-world video databases. The results are published at the CVPR 2011 conference in the paper Actom Sequence Models for Efficient Action Detection (downloadable here). |
|
| More details are available on the projects page. |
| During this project, we aimed at using textual and visual information to extract, from movie databases, some video examples of a query action. The paper Mining visual actions from movies (downloadable here), published at the BMVC 2009 conference, presents our approach and gives some results on the TV-show Buffy the vampire slayer. |
|
| More details are available on the projects page. |
