CHARACTER DISCRIMINATION PROCESSING SYSTEM

• Main Author: IIDA YUKIYASU
• Format: Patent
• Language: English
• Published: 17.02.1984
• Subjects: CALCULATING; COMPUTING; COUNTING; PHYSICS

PURPOSE:To store a word dictionary which is effective for character reading in an OCR and to uses its information for a high-precision read of characters including similar character patterns, by reading a document that an user of the OCR enters words to be read into corresponding to an application through the OCR device, and checking and correcting read results corresponding to the input characters. CONSTITUTION:A scanning part 2 scans on a document 2 and transfers the picture signal of an input character to a recongition part 3. When the input document is a word registration document, a common control part 1 operates a word dictionary generation part 5 to generate a word dictionary on the basis of a character code string of a word on the document for word registration which is sent from the recognition part 3. Then when the input document is a word dictionary specifying document, a word collation processing part 7 is actuated to read a dictionary having a dictionary name entered in the word dictionary specifying document out of a word dictionary storage part 6. Then, when a document to be read is inputted, a candidate character code string for the character string is sent, field by field, from the recognition part 3 to the word collation processing part 7, which collates the word dictionary specified, field by field, with the candidate code to sent out the character code of the most coincident word dictionary to an output line 11.