Robust Visual Tracking using Sparse Principle Component Analysis and Haar-like Features

Object tracking is an important problem with wide ranging applications. The aim of object tracking is to detect object and track its motion in the video. Issues of concern are to be able to map objects correctly between two frames, and to be able to track through occlusion. Factors such as pose variation, illumination change, occlusion, and motion blur make it difficult to develop effective and efficient appearance models for robust object tracking. A new tracking method combining on sparse representation, PCA and Haar-like features is proposed in this paper. Occlusion can be evaluated in PCA based analysis in Bayesian inference framework. Compressed features using sparse representation matching method is used to locate the target object if the level of occlusion satisfies inequality criteria. Most of the unimportant samples are removed before computing the compressed features; hence the computational complexity of proposed algorithm is maintained in optimum level. Experiments show that the proposed method with performs better against illumination change, occlusion and appearance variation, and outperforms several latest important tracking methods in terms of tracking performance.