Compressive tracking combined with sample weights and adaptive learning factor

摘要

The compressive tracking algorithm introduces a compressive sensing theory into the targettracking field andproduces good real-timeperformance.However, the original compressive trackingalgorithm ignores the fact that individual samples make different contributions to the targetand that the learning factor is an empirical value that remains constant when the template isupdated. Therefore, adverse factors (such as noise) and errors can infiltrate into the parametricmodel during the updating of the modelwhen the object is obscured or receives interference fromexternal factors, which will lead to tracking drift. In view of these problems, the weights of samplesare given according to the distance between the sample and the target when training theNaive Bayesian classifier; hence, the stability of the tracking is improved. While the introductionof the Bhattacharyya coefficient is utilized to adjust the learning factor, this can help parametersto self-adapt effectively. Experimental results show that the improved tracking algorithm has abetter adaption to the target appearance variations, illumination changes, occlusion, and so on,and has better robustness than the original algorithm.