Diffusion approximation on neural networks and applications for image processing
Georges‐Henri Cottet · European Consortium for Mathematics in Industry · 1992
We derive reaction-diffusion approximations for simple models of neural networks. On the basis of this formal link we construct a non linear diffusion operator which combined with reaction allows to obtain non trivial asymptotic states. The efficiency of this model in terms of image processing is discussed on test images as well as medical images.