|Correcting errors of the document reader.|
|Correcting documcni layout.|
|Delete a character: double horizontal lines(=) on the target character.|
|Merge two parts into a single character: U pattern to hold the character.|
|Separate a character into two: a separation symbol (upside down of T).|
|Replace a character: vertical line (|) on the character.|
Interpretation and execution of correction marks . The system interprets correction marks and applies corrections to their targets. For example, when it reads a separation symbol, it performs the character extraction process again to the character image data stored in the output file. When, it reads a merge symbol, it performs character recognition to the combined image of two characters designated.
We made a prototype system on a PC and asked several people to try the system. Some of them were surprised by the result since their pencil corrections were reflected to the result. Others said that they prefer editing on PC. At the moment, a set of correction marks to be recognized is too small. Moreover, recognition of correction marks sometimes fails. After solving simple problems revealed in the prototype system, we will evaluate this method more precisely. By this initial prototype, however, we have confirmed that the people who do not want to use a keyboard or PC can accept the method.
Suenaga, Y. ( 1980). A facsimile based text editor, The Transaction of the Institute of Electronics, Information and Communication Engineers, J63-D, 12. 1072-1079(in Japanese).
Bunke, H., Gonin, R. & Mori, D.( 1997). A tool for versatile and user-friendly document correction. Proc.4th ICDAR'97, (Vol. 1) Aug. 1997, 433-438.