| Abstract: | A homogeneous neuron-like medium with close nonlocal coupling between the positive center and the negative surroundings is used for image processing. The input image is transformed into a stationary structure dependent on the form of the coupling function and the type of the element's response to external action. Using the idea of plausible activity structures, we can determine the model parameters needed to extract the features of the medium such as the contour, the lines of definite direction, the ends of lines, the central axes of the image, the negative, the points of intersection of lines, the texture boundaries, and the objects of definite scale as well as the parameters that permit reconstruction of broken lines, smoothing of lines, automatic determination of image scales, and binarization of the initial image by a medium with a tunable coupling function.Institute of Applied Physics, Russian Academy of Sciences, Nizhny Novgorod. Translated from Izvestiya Vysshikh Uchebnykh Zavedenii, Radiofizika, Vol. 37, No. 8, pp. 1053–1061, August, 1994. |
