Areal Multistart Method of Optimization for Image Recognition

摘要

The multistart method of optimization for image recognition tasks has been developed. This method allows receiving an optimal solution in the form of the convergent sequence of areas to area pragmatic (semantic) sufficiency (APSS). The application of the method in the framework of the information-statistical approach will allow the formation of information technologies in visual information processing systems with the possibility of changing the error and speed of procedures based on optimization when the conditions for obtaining visual information change.