| 摘要 |
<P>PROBLEM TO BE SOLVED: To realize an image reading device for reading a document image capable of automatically deciding the language of characters written in an original. <P>SOLUTION: This image reading device 1 is provided with a labeling processing means 3 for grouping black pixel continuous areas as one group from among binary image data read by this image reading device, and for extracting the circumscribed rectangle information of the group, a line extraction processing means 4 for extracting line rectangle information from the position information of the extracted group circumscribed rectangle and an integration processing means 5 for deciding the overlap of the group circumscribed rectangles included in the extracted line rectangle, and for executing overlap integration processing to integrate the overlapped groups into one group. The rate of the number of the group circumscribed rectangles included in the line rectangle prior to the overlap integration processing to the number of the group circumscribed rectangles included in the line rectangle posterior to the overlap integration processing is calculated, and the language of characters written in an original is judged from the difference of the rate. <P>COPYRIGHT: (C)2005,JPO&NCIPI |
