Publication
Constraining human motion for efficient tracking with a particle filter
Conference Article
Conference
CVC Workshop on Computer Vision: Advances in Research & Development (162)
Edition
1st
Pages
1-6
Doc link
http://ccuc.cbuc.cat/record=b4137133~S23*cat
File
Abstract
Particle filters are one of the most commonly used techniques for full-body human tracking. However, given the high-dimensionality of the involved models, the number of required particles make the problem computationally very expensive. To overcome this, we present an action specific model of human postures which eases the process by guiding the prediction step of the particle filter, so only feasible human postures are considered. Thus, this model-based tracking approach samples from a first order motion model only those postures which are accepted by our action-specific model. In this manner, particles are propagated to locations in the search space with most a posteriori information avoiding particle wastage. We show that this scheme improves the efficiency and accuracy of the overall tracking approach.
Categories
computer vision.
Author keywords
motion analysis and recognition, particle filters
Scientific reference
I. Rius, C. Fernández, M. Mozerov and J. Gonzàlez. Constraining human motion for efficient tracking with a particle filter, 1st CVC Workshop on Computer Vision: Advances in Research & Development, 2006, Bellaterra, Espanya, pp. 1-6, 2006, UAB, Bellaterra, Espanya.
Follow us!