View Publication

Shape Time Discriminative Classification of Video Objects

Li Cheng
Baochun Bai, Department of Computing Science, University of Alberta
Cheng Lei, Department of Computing Science, University of Alberta
Dale Schuurmans, AICML
Shaojun Wang, Dept of Computing Science

We propose a discriminative approach to non�rigid video objects classi� fication. Our goal is to recognize actions of the objects that appear in a video sequence, based on its shape time dynamics. This is achieved by exploiting the temporal action correlations under large�margin struc� tured classification framework. Further, it leads to an algorithm that can naturally utilize non�vectorial shape representation and matching tech� niques, which are difficult to incorporate into the commonly used hidden Markov models (HMMs). The proposed approach is verified on indoor video sequences for action classifications of human subjects.

Citation

L. Cheng, B. Bai, C. Lei, D. Schuurmans, S. Wang. "Shape Time Discriminative Classification of Video Objects". 2005.

Keywords:	discriminative, machine learning
Category:

BibTeX

@incollection{Cheng+al:05,
  author = {Li Cheng and Baochun Bai and Cheng Lei and Dale Schuurmans and
    Shaojun Wang},
  title = {Shape Time Discriminative Classification of Video Objects},
  year = 2005,
}

Last Updated: May 03, 2006
Submitted by Lori-Ann Peredery

Not Logged In

PapersDB

Shape Time Discriminative Classification of Video Objects

Citation

BibTeX