Robust motion estimation using spatial Gabor filters

This paper presents a new algorithm for motion estimation. It combines Gabor filter decomposition and robust least squares estimation in a multiresolution framework. Spatial Gabor filter bank provides a multichannel decomposition of frame sequence. Then, applying the brightness constancy constraint on each channel between two consecutive frames, we obtain an overdetermined system of velocity equations at each pixel. In order to be robust to outliers, this overdetermined system is solved using a robust least squares technique. We have used a multiresolution framework in order to manage large and small displacements. Performances of our algorithm are tested on synthetic and real sequences, and are compared with other techniques.