FIELD: information technology.
SUBSTANCE: method of resolving conflicting output data from an optical character recognition system (OCR), which enables conversion of bitmap documents into text in computer codes as output data. Output data of the OCR system include at least first and second characters included in the list of possible candidates for a copy of the same selected character specimen from the bitmap document. Resolution of conflicting output data is carried by performing steps on which locations of differences between candidate characters in graphic form are identified, and information on the locations is used to identify corresponding positions in the selected character specimen. Based on a correlation technique, information on location is used to select the correct candidate character as identification of the selected character specimen.
EFFECT: reducing uncertainties associated with selection of correct candidate characters among several candidate characters.
20 cl, 31 dwg, 2 tbl
