A recurrent neural network for image flow computation

The image flow computation in dynamic image processing can be formulated as a minimization of functionals. The authors show that this formulation can be solved by a recurrent neural network. They start with the Euler necessary condition and natural boundary condition, then derive a set of difference equations. Based on the analysis of the equations, a recurrent neural network is proposed for solving image flow. Experiments on synthetic and real laboratory image data were performed. The proposed network can be implemented in hardware