Domain Adaptation in the Context of Sport Video Action Recognition
FarajiDavar, N, deCampos, TE, Windridge, D, Kittler, J and Christmas, W (2011) Domain Adaptation in the Context of Sport Video Action Recognition In: NIPS 2011 Domain Adaptation Workshop, 2011-12-17 - 2011-12-17, Sierra Nevada, Spain.
FarajiDavar-DA-2011.pdf - Accepted Version
Available under License : See the attached licence file.
We apply domain adaptation to the problem of recognizing common actions between differing court-game sport videos (in particular tennis and badminton games). Actions are characterized in terms of HOG3D features extracted at the bounding box of each detected player, and thus have large intrinsic dimensionality. The techniques evaluated here for domain adaptation are based on estimating linear transformations to adapt the source domain features in order to maximize the similarity between posterior PDFs for each class in the source domain and the expected posterior PDF for each class in the target domain. As such, the problem scales linearly with feature dimensionality, making the video-environment domain adaptation problem tractable on reasonable time scales and resilient to over-fitting. We thus demonstrate that significant performance improvement can be achieved by applying domain adaptation in this context.
|Item Type:||Conference or Workshop Item (Paper)|
|Additional Information:||This project was sponsored by the EPSRC/UK grant EP/F069421/1 (ACASVA) and EU PASCAL2. This is the preprint of a paper that appears in NIPS Domain AdaptationWorkshop, Dec 2011|
|Divisions:||Faculty of Engineering and Physical Sciences > Electronic Engineering > Centre for Vision Speech and Signal Processing|
|Depositing User:||Symplectic Elements|
|Date Deposited:||23 May 2012 12:07|
|Last Modified:||23 Sep 2013 18:58|
Actions (login required)
Downloads per month over past year