| 摘要 |
For the problem of blurring at a boundary in a scanned illustration, a method of separating and vectorizing color regions accurately is realized by identification of a labeled region corresponding to the blurring and combining processing with another neighboring region. An image input unit, a color region dividing unit dividing an input image into the color regions, a labeling unit providing an individual label to the color region, a labeled region shape analysis unit obtaining labeled region shape information including information representing a shape feature of a labeled region which is the color region provided with the label, a micro-region determination unit determining whether the labeled region provided with the label is a micro-region or not according to the labeled region shape information, and a micro-region combining unit combining the region determined to be the micro-region with a neighboring region, are comprised.
|
