Title :
Pattern Recognition Using Autocorrelation
Author :
Horwitz, L.P. ; Shelton, G.L., Jr.
Author_Institution :
IBM Mohansic Lab. Res. Ctr., Yorktown Heights, N.Y.
Abstract :
A class of techniques for character recognition is described. These techniques are characterized by the property that the only parameters of the input which are used are those which are independent of the position of the character; that is, these techniques are registration invariant. In Section I of this paper, we describe the registration invariant mathematical formalism which underlies these techniques, and in Section II we describe in more detail the physical realization of several recognition schemes based on these concepts. Some results from our computer simulation of these procedures are given.
Keywords :
Autocorrelation; Character recognition; Computer simulation; Data processing; Design engineering; Geometrical optics; Optical character recognition software; Optical diffraction; Pattern recognition;
Journal_Title :
Proceedings of the IRE
DOI :
10.1109/JRPROC.1961.287787