| 摘要 |
<p>A document is scanned by an image scanner and image data representing the image of the document is stored in an image storage means. Rectangles contacting and surrounding outer boundaries of each image of plural character rows in the image storage means are generated, and the positions of four edges of the rectangle in XY coordinates of the image storage means are detected. A size of a rectangle is calculated based upon the detected positions of the edges, and the size of each rectangle is compared with a expected size range for the characters and symbols to be recognized. The positions of the rectangles falling into the size range are stored in a first table as the position data wherein the position data of the rectangles of the characters and symbols over the plural character rows are arranged in the first list in the order from a rectangle at one end to a rectangle at the other end along the direction of the X axis of the XY coordinates. The position data of the rectangles in the first list are sequentially fetched in the arranged order to detect a size of each rectangles to determine an average size of all rectangles stored in the first list. Again, the position data of the rectangles in the first list are sequentially fetched in the arranged order to find out a first rectangle falling into a size range settled based upon the average size. The fetch operations are continued to find out a second rectangle having a bottom left corner located within predetermined distances in the X and Y directions from a bottom left corner of the first rectangle. The fetch operations are continued to find out a third rectangle having a bottom left corner located within the predetermined distances in the X and Y directions from the bottom left corner of the second rectangle. The operations continues to find out a predetermined number of rectangles in one character row satisfying the condition. When the predetermined number of the rectangles have been found, a skew of the character row in the XY coordinates is calculated based upon the positions of the bottom left corners of these rectangles, and this detected skew is treated as the skew of the document.</p> |
