JAPANESE/ENGLISH DISCRIMINATING METHOD FOR DOCUMENT IMAGE AND RECORDING MEDIUM

摘要

PROBLEM TO BE SOLVED: To accurately identify Japanese and English at high speed and to identify both the languages even concerning a range to be identified for each character area and each page unit.;SOLUTION: After an input document image 101 is reduced 102, a black pixel connection component is extracted 103, and the character area is generated 104 by merging these components. Concerning the generated character area, based on the length of the connection component, a Japanese/English discriminating means 105 classifies that component and based on the accumulated value of the classified result, whether it is a Japanese area or an English area is discriminated.;COPYRIGHT: (C)2006,JPO&NCIPI