DMCA
Multi-Task Sparse Learning with Beta Process Prior for Action Recognition
| Citations: | 1 - 0 self |
BibTeX
@MISC{Yuan_multi-tasksparse,
author = {Chunfeng Yuan and Weiming Hu and Guodong Tian and Shuang Yang and Haoran Wang},
title = {Multi-Task Sparse Learning with Beta Process Prior for Action Recognition},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
In this paper, we formulate human action recognition as a novel Multi-Task Sparse Learning(MTSL) framework which aims to construct a test sample with multiple fea-tures from as few bases as possible. Learning the sparse representation under each feature modality is considered as a single task in MTSL. Since the tasks are generated from multiple features associated with the same visual in-put, they are not independent but inter-related. We intro-duce a Beta process(BP) prior to the hierarchical MTSL model, which efficiently learns a compact dictionary and infers the sparse structure shared across all the tasks. The MTSL model enforces the robustness in coefficient estima-tion compared with performing each task independently. Besides, the sparseness is achieved via the Beta process for-mulation rather than the computationally expensive l1 norm penalty. In terms of non-informative gamma hyper-priors, the sparsity level is totally decided by the data. Finally, the learning problem is solved by Gibbs sampling inference which estimates the full posterior on the model parameters. Experimental results on the KTH and UCF sports datasets demonstrate the effectiveness of the proposed MTSL ap-proach for action recognition. 1.
Keyphrases
action recognition multi-task sparse learning beta process prior beta process for-mulation non-informative gamma hyper-priors beta process visual in-put test sample hierarchical mtsl model learning problem model parameter full posterior multiple fea-tures mtsl ap-proach feature modality mtsl model human action recognition sparse structure expensive l1 norm penalty multiple feature ucf sport datasets coefficient estima-tion novel multi-task sparse learning sparsity level single task sparse representation experimental result
