Research on Target Tracking Algorithm Based on Variance Sampling for Fast Sparse Representation

摘要

The Sparse Representation tracking algorithm under particle filter framework has many problems, such as large number of particles and complexity of$l_{1}$norm minimization calculation. To refine the issues mentioned above, a fast sparse representation target tracking algorithm based on variance sampling is proposed. The algorithm first uses the variance estimation to optimize the distribution of sampling particles in the motion estimation stage; Then, the discriminant objective function is solved by using$l_{2}$norm instead of$l_{1}$norm, furthermore, the improvement of the measurement form of the reconstruction error is made to enhance the sparsity of the$l_{2}$norm; Finally, online dictionary learning (ODL) algorithm to update the template dictionary online. The experimental results show that the proposed algorithm can greatly improve the tracking efficiency under the premise of ensuring stable tracking and accuracy.