Minimization of Visual Information

Abstract

One of the important issues in the problem of automatic recognition of visual images is the question of reducing the amount of input information into the CM (Computing Machine) of the initial information (the number of image discretization elements). The degree of discretization of pictorial (multi-gradation) image of objects for normal operation of recognition algorithms can be hundreds of thousands of discretization elements. Such a volume of data, even after their compression during encoding, is difficult to insert into the operative memory of modern computers. When using the same long-term memory of the machine dramatically increases the time of the task. Therefore, not all information about objects should be entered into the memory of the machine, but only its most informative part.