Artificial intelligence

Abstract: A grid-density-center method of the reduction of the teaching selections in the recognition systems is proposed. It is based on coverage of character space and on finding the unique object of a cage as object of new teaching selections. Principle of forming of the grid and methods of the construction of the objects of brief teaching selection are offered. For calculation of the efficiency of the offered method a comparative experimental analysis is conducted with the known methods. The analyses have shown that the method increases accuracy the classification and decrease the length of the teaching selections.

Keywords:

References: