Integral Channel Features for Particle Filter Based Object Tracking

In this paper, we propose an object tracking algorithm Based on particle filter using integral channel features. Integral channel features are the extension of features which can be computed using the integral image of multiple image channels. They combine diversity of information and high computational efficiency. In this algorithm, two kinds of integral channel features (the gray and the gradient magnitude) are combined in particle filter framework. The appearance model is part Based, which makes it robust to occlusions. We test the proposed method over three challenging sequences involving partial occlusions, drastic illumination changes and similar-color interference. Our method shows excellent performance in comparison with three previously proposed trackers.