FORMULA RECOGNIZING DEVICE AND FORMULA RECOGNIZING METHOD

摘要

PROBLEM TO BE SOLVED: To realize a formula structure analysis hardly subject to errors in character recognition results and special font characters. SOLUTION: A formula detection part 114 has a line center zone detection portion 201 and a suffix determination portion 202 for determining a suffix structure. In the line center zone detection portion 201, a histogram for the coordinates of central portions of a plurality of characters belonging to a line is drawn for every line including numerical formulae and its maximum peak value is determined as a center zone coordinate which forms a center zone of the line. In the suffix determination portion 202, the character recognition results of the individual characters are never used and the suffix structure is determined for every character in accordance with vertical relationship between the circumscribed rectangle of each character in the line and the center zone coordinate in the line. COPYRIGHT: (C)2003,JPO