Underlying neural computations for some visual phenomena (Q1120515)

This work continues previous investigations of the same (or some of the) authors, see e.g. ibid. 54, 21-28 (1986; Zbl 0585.92012), and ibid. 53, 239-245 (1986; Zbl 0587.92011), on a continuous neural model applied to modelling of vision phenomena. The notion of receptive field for the continuous model is introduced and a receptive field of a circular form is considered. Numerical simulations are carried out using a formal discretization of the model. It is shown that a neural network of 33\(\times 33\) neurons with the defined type of spatio-temporal receptive field interacts to produce a number of known psychophysical effects like: masking, Mach bands, constrast, meta-contrast, Herman grid illusions.