Authors: Adam Gonczarek Jakub M Tomczak
Publish Date: 2016/02/02
Volume: 27, Issue: 2, Pages: 275-286
Abstract
In this paper we investigate articulated human motion tracking from video sequences using Bayesian approach We derive a generic particlebased filtering procedure with a lowdimensional manifold The manifold can be treated as a regularizer that enforces a distribution over poses during tracking process to be concentrated around the lowdimensional embedding We refer to our method as manifold regularized particle filter We present a particular implementation of our method based on backconstrained gaussian process latent variable model and gaussian diffusion The proposed approach is evaluated using the reallife benchmark dataset HumanEva We show empirically that the presented sampling scheme outperforms samplingimportance resampling and annealed particle filter procedures
Keywords: