Object tracker using sparse prototypes and annealed particle filter

Object tracking is a challenging problem as it entails learning an effective model to account for appearance change caused by the presence of occlusions, pose variations, background clutter and drastic illumination changes. In this paper, we propose a novel object tracking algorithm with sparse prototypes and the annealed particle filter (SPAPF), which exploits both recent sparse prototypes with classic annealed particle filter schemes for learning robust tracking models. We introduce annealed particle filter into the recent sparse prototypes for object tracking, and develop a novel algorithm to represent an object by sparse prototypes that account explicitly for noise and the annealed particle filter, which can avoid generating invalid particles corresponding to impossible target object. Results on challenging image sequences demonstrate that the proposed tracking algorithm performs better in comparison with the state-of-the-art sparse prototypes tracking algorithm.