摘要 |
PROBLEM TO BE SOLVED: To precisely discriminate between a character and a dot character in document image data and to use processes suitable for them respectively. SOLUTION: An image region separation section 22 processes input RGB image data to generate a signal wh indicating a white region by a white region extraction section 222, a signal ht indicating a dot detection result by a dot separation section 223, a signal edge indicating an edge detection result of a character etc., by an edge extraction section 224, and a signal hted indicating an edge detection result of a dot character and a signal iro indicating a result of a chromatic/achromatic decision from a color decision section 226 by a dot edge extraction section 225, and outputs a C/P signal (respective regions of a pattern, a character, a dot character, and a dot) representing a result (image type) of image region separation and a B/C signal (representing a chromatic region or an achromatic region) based upon those signals. COPYRIGHT: (C)2006,JPO&NCIPI
|