Sequential quasi-Monte Carlo filter for visual object tracking

This paper investigates an object tracking algorithm using a sequential quasi-Monte Carlo (SQMC) filter combined with covariance features. Covariance features are used not only to model target appearance, but also to model background. By incorporating the dissimilarity of target and background into the SQMC filter, the proposed SQMC filter improves the accuracy of the particle weight. A target model update strategy using the element of Riemannian geometry is proposed for the variation of the target appearance. Comparison experiments show that the proposed algorithm can successfully track the object in the presence of appearance changes, cluttered background.