Neuronal approach for motion estimation in frequency domain

The analysis of motion in image sequences is a field of research still open. The development of new methods is guided by two objectives: the first is to improve the results obtained from the qualitative point of view; the second goal is to minimize the calculation time to meet the real-time applications. Our work is based on the application of neuronal motion estimation approach called MENA in the frequency domain. This choice is guided primarily by the qualitative results of this method and also by the strength of the frequency representation of the processed data. This view allows us also to generalize the MENA method that can benefit from the contributions of the different approaches to motion estimation.