The paper proposes a method of calculating the amount of visual information in an image. The method is based on the model of local analysis in the early vision. We are talking about the information perceived by an observer under certain conditions, not about the information contained in the raster image per se. Simple striate cortex neurons respond to the simple visual features of image, for example, the bright stripes in a dark environment, and the brightness gradients. To evaluate the amount of information in the image, it is possible to take the ratio of the number of activated neurons to the total number of neurons in a model. It is necessary to consider the angular size of the image to calculate the value of the estimate.

Keywords: Simple visual feature, raster image, information, entropy, redundancy, simple striate neuron, Gabor function, computer simulation

In this report, we propose an approach to the automatic extraction of such components of the image, that are most useful for human observers to solve the recognition task. This approach is based on the concepts of psychophysiological mechanisms of grouping the simple features at the preattentive stages of visual processing. A method is based on digital image processing by a software implemented model "filter-rectify-filter". The extraction of the most informative areas of the photograph is based on finding local maxima of the model "outputs". The calculation results indicate that the areas of the image containing spatial modulations of the local features may be very useful for the recognition of faces.

Keywords: Second order visual mechanisms, image filtering, regions of interest, face identification, computer simulation

This paper outlines the implementation of computer model of second-order visual mechanisms that are specific to spatial contrast modulations, and represents the results of testing the model. Computer simulation is based on "Filter-Rectify-Filter" model with some important additions. The inhibitory parts of receptive field of simulated mechanism are non-selective to orientation and spatial frequency of texture elements. Texture modulated by orientation or spatial frequency does not reduce flank inhibition and does not activate the mechanism. Unmodulated texture  also does not cause activation of the filter because the flanking inhibition will be equal to or exceed the excitation. However, increasing the contrast in the center and reduction it on the flanks leads to the activation of the second order mechanism.

Keywords: Second order visual mechanisms, image filtering, texture, spatial modulation, computer simulation