A MATHEMATICAL MODEL OF PRIMARY INFORMATION PROCESSING IN VERTEBRATE VISUAL SYSTEM
视觉系统初级信息加工的一种数学模型——<Ⅰ>模型的空间性质

If the vertebrate visual system is considered as linear and homogeneous for image processing in temporal and spatial domains, its kernal function is the key to describing the properties of the system. In this paper, we take the extended Gabor function (EG) as a kernel function for primary information processing of visual system. The so-called extended Gabor function is obtained as the original Gabor elementary signals are extended from one independent variable to three ones (including one variable in time and other two in space),We suggest that EG take two major forms,one is in the isotropic system, the other is in the non-isotropic system.If the parameters in EG function are chosen properly, some main types of receptive field discovered in electrophysiological experiments (for example, Kuffler's and Hubel & Wiesel's receptive fields) can be described by this model qualitatively. Another types of receptive fields can be predicted from this model. If logarithmic transform is introduced in spatial frequency domain, the modulation transfer f unction (MTF) of human vision can be synthesized by taking Fourier transform of the EG functions.