| 摘要 |
A line orientation determining unit (1) determines whether the line orientation of an input document image is vertical or horizontal. A character image extracting unit (2) extracts a character image from the input document image, and a character orientation detecting unit (3) detects whether a rotation is required, if so finds the rotational angle, and detects whether a mirror image flipping is needed, according to the determination made by the line orientation determining unit. A document image correcting unit (4) corrects the input document image to the proper orientation based on the result of the detection made by the character orientation detecting unit (3). This allows correction of a scanned image of a document such as an English document, the text of which is horizontally written in normal cases, scanned in any of upward, downward, right, and left directions, for example, with the use of a hand-held scanner. |
