Minimax Monte Carlo object tracking

We propose a new approach for visual object tracking based on a combined method of minimax estimator and sequential Monte Carlo filtering. The proposed approach adopts a minimax strategy in the standard particle filtering framework for the problem. Particle filtering is based on probabilistic methodology, while a minimax estimator belongs to deterministic approaches. Experiments show outperforming results of the proposed approach compared to the standard particle filtering in terms of tracking accuracy. We also investigate the computational complexity of the proposed algorithm in terms of elapsed processing time. In this paper, we focus on the particle filtering framework only for the performance comparison between the two methods.