Methods are in-process examined classifications, applied at the compression of file with the bitonality image of phototypograph got scan-out or digital photographing. The special attention applies on the in-use here measures of distinction of two images of characters, abstracted from the image of text. These measures of distinction allow with one or another degree of confidence to count characters on the compared images or consilient, or different. For the algorithms of classification known to date, including the known algorithm of JB2 well, quantitative descriptions of classification – number of classes, got these algorithms for the image of standard page of text are resulted. What less than it is a number, quality of classification is considered that higher, because gives the best compression of file with the image of text. The methods of acceleration of algorithms, classifying the images of characters, and increases of easy-to-readness of the text recovered after the compression of image are considered also.
image of text, methods of classification, compression of data