Формирование отношений аффинных базисов при распознавании по структурным признакам

Abstract

This work is dedicated to the solving of such problem as comparing visual objects based on different characteristics. Affine-transformed images with defined features had been taken as investigated objects. The modification of geometric hashing method was suggested. The number of experiments were held for definition of the speed of working and the results accuracy. The optimal limit of comparing with the hash-table was reached in the empirical way.