Gradient sparsity for piecewise continuous optical flow estimation

This paper introduces a new sparse model for robust and reliable optical flow estimation. We show that the sparsity is directly related to gradient fields of optical flow. According to theory on sparse signal recovery, we rigorously formulate the optical flow estimation as an ℓ₀ optimization problem. Considering the piecewise continuous nature of motion, the basic optical flow constraint is regularized by an ℓ₁ norm. Then, with convex relaxation, the solution is obtain via the ordinary or reweighted ℓ₁ optimization approach. Experimental results show the proposed method performs better than traditional methods and deals well with the discontinuities on motion boundaries.