Q-processors for real-time image processing

摘要

In the paper was considered increase of processors performance for solution of image recognition and images analysis (IRIA) problems is very actual for applied mathematics and related branches of science and technology, where the progress is achieved by real-time scale (RTS). The paper presents the main developments of the authors in a such vital and immortal for functioning of Q-processors problems as organization, comparative analysis of QQ-transformation (QQT) structure an neurolike structure, completeness and sufficient of Q-characteristics decomposition, algorithms of image restoration from the set of its Q-characteristics, noise-immunity and information content of QQT, result of comparative mathematical modeling, etc. Which are suggested in a number of papers, however, the analysis, performed in our paper, of certain problems dealing with methodology of generalized Q-transformation. Extreme simplicity of hardware realization provides achievement of maximum possible parallelism of data processing at minimum values of energy consumption and dimension- weight characteristics