Parallel processing of figure-ground separation in random dot kinematogram with optical flow discretization

Parallel processing is an important factor for fast image processing in the computer vision as well as human visual systems. From this point of view, this paper treats the figure-ground separation problem in the random dot kinematogram. To achieve this separation, the optical flow is discretized in the five vectors, each of which the flag is assigned to indicate its correct direction in each pixel. The recognition process is described as the flag dynamics, which is given as the gradient system of the global potential functional. The performance is examined with computer simulations.