| 摘要 |
<p>PURPOSE: To accurately detect the tilt of an input document without user's intervention by detecting the tilt quantity of inputted image information according to projections detected from two determined partial image areas. CONSTITUTION: An image which is optically read by a scanner 108, etc., is inputted. The inputted image is stored in a RAM 103. To find the tilt quantity of a document in the input image, an inspection area is acquired in the input image. The image in the obtained inspection area is analyzed to decide the document direction in the area. The tilt quantity of an area wherein the document direction is judged to be lateral is found by a tilt quantity detecting method for lateral writing and the tilt quantity of an area which is judged to be a longitudinal writing area, on the other hand, is found by a tilt quantity detecting method for longitudinal writing. Thus, the tilt quantity is found and the input image is so deformed as to remove the tilt of the document. Namely, the image is rotated around the center of gravity of the input image by the tilt quantity in this case.</p> |
