| 摘要 |
PURPOSE:To correct a dictionary for correction without necessity for a user to correct an uncorrected part based on a document coded in advance. CONSTITUTION:This method is composed of a document file 1 for storing coded documents, character font file 2 for storing character fonts, image file 3 for storing document images, dictionary 4 for correction to store the feature amounts of characters, dictionary 5 for correction to store the rules of character recognition error correction, image generating part 7 for turning a document to image data, character recognizing part 8 for performing rectangle segmentation and character recognizing processing, recognition correcting part 9 for performing correction based on the rules in the dictionary for correction, dictionary maintaining part 10 for registering or deleting the rules, control part 6 for controlling various kinds of processing such as display and input, display means 11 for promoting the input of layout, and input means 12 for inputting the layout. |
