| 摘要 |
<P>PROBLEM TO BE SOLVED: To solve the problem in which: an appropriate model color cannot be extracted if a multi-valued image is largely reduced before a model color for characters is extracted so that a file size is drastically reduced; and a large memory is required for model color extraction processing if an attempt is made to extract the model color before the input multi-valued image is reduced. <P>SOLUTION: A multi-valued image is reduced to generate a first reduced multi-valued image, and a model color of a character region is extracted from the first reduced multi-valued image. Then, the first reduced multi-valued image is reduced to generate a second reduced multi-valued image, and the second reduced multi-valued image is compressed to generate a first compression code. Also, a partial binary image corresponding to the character region is compressed to generate a second compression code. Then, compressed data is output which includes the first compression code, the second compression code, position information of the character region, and model color information of the character region. <P>COPYRIGHT: (C)2013,JPO&INPIT |
