Body pose based pedestrian tracking in a particle filtering framework
M. Rahman, Jesús Martínez del Rincón, Jean‐Christophe Nebel, Dimitrios Makris · Research Repository (Kingston University London) · 2013
A novel body pose based human tracking model is proposed for pedestrian tracking.This work investigates the challenges of reliable pedestrian tracking and proposes an improved model under challenging environments.Specifically, it claims that it is useful to exploit the curvature information of different body poses in tracking framework to overcome general tracking problems.In this paper different body pose detectors are combined as a useful feature for tracking.Performance has been evaluated in a rich evaluation framework.Result shows that poselet based features are more suitable for tracking than just detecting the person over the frames.