| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide a character recognition device which discriminates between original space characters which exist in a document and are recognized and space characters which do not exist in the document and are erroneously recognized because preceding or succeeding characters having relatively small character widths, to delete only the erroneously recognized space characters when performing character recognition of image data including European characters in a fixed-width font. <P>SOLUTION: A character recognition device 10 recognizes characters from a read document and takes space characters as delimiters to correct character strings of character recognition results per word. The character recognition device 10 includes: a circumscribed rectangle forming unit 17 which forms a circumscribed rectangle of each of recognized alphabetic character strings; a fixed-width font determination unit 19 which determines whether each character string is in a fixed-width font or not on the basis of distances between center lines in a breadthwise direction of adjacent circumscribed rectangles; an excess space character determination unit 20 which determines space characters in the character string to be excess, on the basis of the fact that, in the case of a fixed-width font, widths of the space characters are smaller than a prescribed width; and a deletion unit 21 which deletes the space characters determined to be excess, from the character string. <P>COPYRIGHT: (C)2013,JPO&INPIT |
