PCA-based adaptive particle filter for tracking

The particle filter is a popular tool for visual tracking. Traditionally, the number of particles used is typically fixed, and the motion model is simply a random walk with fixed noise variance. All these factors make the visual tracker unstable. To stabilize the tracker and guarantee the real-time tracking, an adaptive particle filter algorithm which estimates the motion model parameters using principal component analysis (PCA), and adaptively selects the number of particles and the motion model parameters are proposed in this paper. Experimental results indicate that the proposed method enhances performance of the vision tracking based on particle filter.