Bilinear Programming for Human Activity Recognition with Unknown MRF Graphs
Authors:
- Zhenhua WangZWZhenhua Wang
- Qinfeng ShiQSQinfeng Shi
- Chunhua ShenCSChunhua Shen
- Anton van Den HengelAHAnton van Den Hengel
- 1 June 2013
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10636919,p. 1690-1697
Abstract
Markov Random Fields (MRFs) have been successfully applied to human activity modelling, largely due to their ability to model complex dependencies and deal with local uncertainty. However, the underlying graph structure is often manually specified, or automatically constructed by heuristics. We show, instead, that learning an MRF graph and performing MAP inference can be achieved simultaneously by solving a bilinear program. Equipped with the bilinear program based MAP inference for an unknown graph, we show how to estimate parameters efficiently and effectively with a latent structural SVM. We apply our techniques to predict sport moves (such as serve, volley in tennis) and human activity in TV episodes (such as kiss, hug and Hi-Five). Experimental results show the proposed method outperforms the state-of-the-art.Keywords
This publication has 12 references indexed in Scilit:
- Human activity analysisACM Computing Surveys, 2011
- Learning structural SVMs with latent variables2009
- Learning realistic human actions from movies2008
- On Space-Time Interest PointsInternational Journal of Computer Vision, 2005
- Histograms of Oriented Gradients for Human Detection2005
- Recognizing human actions: a local SVM approach2004
- A method for human action recognitionImage and Vision Computing, 2003
- The Concave-Convex ProcedureNeural Computation, 2003
- On implementing a primal-dual interior-point method for conic quadratic optimizationMathematical Programming, 2003
- Computability of global solutions to factorable nonconvex programs: Part I — Convex underestimating problemsMathematical Programming, 1976