A new algorithm for optical flow computation is presented using neural networks. The computation procedure consists of three stages: estimation of the parameters of the neural network model

dynamic measurement of the perpendicular velocity components of the contour and optical flow computation. The parameters are estimated by comparing the energy function of the neural network with a constrained error function of motion. The nonlinear optical flow computation method is then carried out iteratively by using a dynamic algorithm to minimize the energy function simultaneously with the dynamic measurement of the perpendicular velocity components by a dynamic procedure. Some factors affecting the convergence property of the neural network are discussed through the simulation results.