BIROn - Birkbeck Institutional Research Online

Action classification using a discriminative non-parametric hidden Markov model

Raman, Natraj and Maybank, Stephen J. and Zhang, Dell (2013) Action classification using a discriminative non-parametric hidden Markov model. In: Vuksanovic, B. and Zhou, J. and Verikas, A. (eds.) Sixth International Conference on Machine Vision (ICMV 2013). SPIE Proceedings 9067. International Society for Optics and Photonics. ISBN 9780819499967.

[img] Text
ActionClass_DiscNPHMM_ICMV.pdf - Author's Accepted Manuscript
Restricted to Repository staff only

Download (481kB) | Request a copy
[img]
Preview
Text
9291.pdf - Published Version of Record

Download (331kB) | Preview

Abstract

We classify human actions occurring in videos, using the skeletal joint positions extracted from a depth image sequence as features. Each action class is represented by a non-parametric Hidden Markov Model (NP-HMM) and the model parameters are learnt in a discriminative way. Specifically, we use a Bayesian framework based on Hierarchical Dirichlet Process (HDP) to automatically infer the cardinality of hidden states and formulate a discriminative function based on distance between Gaussian distributions to improve classification performance. We use elliptical slice sampling to efficiently sample parameters from the complex posterior distribution induced by our discriminative likelihood function. We illustrate our classification results for action class models trained using this technique.

Metadata

Item Type: Book Section
Additional Information: Sixth International Conference on Machine Vision (ICMV 2013) - April 16th 2013, London, UK Copyright 2013 Society of Photo Optical Instrumentation Engineers (SPIE). One print or electronic copy may be made for personal use only. Systematic reproduction and distribution, duplication of any material in this paper for a fee or for commercial purposes, or modification of the content of the paper are prohibited.
Keyword(s) / Subject(s): action classification, depth image, HDP-HMM, discriminative, elliptical slice sampling
School: Birkbeck Faculties and Schools > Faculty of Science > School of Computing and Mathematical Sciences
Research Centres and Institutes: Birkbeck Knowledge Lab
Depositing User: Professor Stephen Maybank
Date Deposited: 06 Mar 2014 09:10
Last Modified: 09 Aug 2025 08:40
URI: https://eprints.bbk.ac.uk/id/eprint/9291

Statistics

6 month trend
0Downloads
6 month trend
0Hits

Additional statistics are available via IRStats2.

Archive Staff Only (login required)

Edit/View Item
Edit/View Item