Semi-local variational optical flow estimation

Global variational methods for optical flow estimation usually suffer from an over-smoothing effect. We propose a semi-local estimation framework designed to integrate and improve any variational method. The idea is to implicitly segment the minimization domain into coherently moving windows. In a first time, local variational estimations are performed in overlapping candidate square regions. Then, a global dis-crete optimization, non subject to the over-smoothing intro-duced by variational approaches, selects the optimal window for each pixel. Experimental results show an increasing of the sharpness of discontinuities and a significant improvement of global registration errors compared to the results of the base-line global variational method. Index Terms — Optical Flow, variational methods, mini-mization domain, motion segmentation, discrete optimizatio