| 摘要 |
PURPOSE:To detect character lines with high accuracy and high efficiency even with a document having an optional character size and optional line space, by presuming the character size and the line space and setting automatically the size of a belt-shaped area at the optimum value based on those presumed character size and line space. CONSTITUTION:A character size/line space presuming means 2 inputs document picture data to obtain the projection of a belt-shaped area in the data and presumes the character size and the line space. A belt area width deciding means 3 sets automatically the size of the belt-shaped area at the optimum value based on said character size and line space. A partial line detecting means 4 and a partial line unifying means 5 divide a document picture into the belt- shaped areas orthogonal to the character line direction in the optimum width to obtain a partial line for each belt-shaped area and connect the partial lines together to the same line for detection a character line. Thus the character lines can be detected with high accuracy and high efficiency even with various documents of different character sizes and line spaces.
|
