Visual tracking with genetic algorithm augmented logistic regression

In this paper, a genetic algorithm (GA) augmented logistic regression tracker is proposed. We enhance our tracker in three aspects. Firstly, a novel concept of intelligent motion model based on GA and particle filter is proposed to handle the partial occlusion, object drift and fast object motion changes during tracking. Secondly, the powerful and efficient features including FHOG and Lab are integrated to further boost the tracking performance. Thirdly, mechanism of dynamic update and choice mechanism of positive and negative templates are introduced to better adapt to the appearance changes. Extensive experimental results on the Object Tracking Benchmark dataset show that the proposed tracker performs favorably against state-of-the-art methods in terms of accuracy and robustness.