摘要 |
PROBLEM TO BE SOLVED: To provide an image region separating method for binary image capable of highly accurately separating a black character region mixing in or adjacent to a dot region from the dot region. SOLUTION: Using a boundary point in image data as a start point, a boundary line is traced in a predetermined rectangular region using the start point as a center point. A boundary line with a closed region existing in the rectangular region is removed to acquire a first mask image, and expansion/shrinkage processing is applied to this mask image to acquire a second mask image. Using a boundary point in the mask image as the start point, a boundary line is traced in a predetermined rectangular region using the start point as a center point, and a closed region existing in the rectangular region is removed from or added to the mask image to acquire a third mask image. Referring to the mask data pixels of this mask image and the black pixels of an original image, if pixels, the outermost pixels in the mask image, adjacent to the pixels of the original image corresponding to the boundary point are black pixels, an pixel corresponding to the black pixel in the mask image is used as an expansion point, and the expansion point is traced and added to the mask image. COPYRIGHT: (C)2005,JPO&NCIPI
|