This article is cited in 4 papers

Analysis of method of character recognition, based on polynomial regression

M. B. Gavrikov, N. V. Pestryakova, A. V. Uskov, V. V. Farsobina


Abstract: Present work contains new results of analysis of method at recognition of handprinted figures on base, coinciding with base of education. Materials concern to features of recognition and mutual location of ‘average’ rasters (or polynomial vectors) of different symbols, and also (in context with them) images of symbols. Generalization of got datum and shown more early monotonous decrease of average estimation of recognition of images of symbol on measure ‘removing’ is executed from 'average' polynomial vector of this symbol. Notions are entered finely-, average- and large-scale phenomena at description of given recognition method. Presence of structures is shown on average- and large-scale levels.