False recognition character table, false recognition character table creation method, character string search device, character string search method, and character string search program

摘要

PROBLEM TO BE SOLVED: To appropriately perform a character string search for text data as an OCR result at low cost even if the printing form of the OCR target image is different without requiring visual confirmation or continuous dictionary update. Allows you to search. SOLUTION: Characters are recognized by scanning a character image printed with OCR target characters with a scanner and performing OCR processing by changing the printing form specified by a printer, a font, and a paper type in various ways. Based on the result of this OCR processing, an erroneous recognition character table that associates erroneously recognized characters with each of the erroneously recognized OCR target characters and also associates the print form (printer, font, paper type) of the character image. To create. When searching for a character string in the text data as an OCR result, the character registered in the erroneous recognition character table in the input character string is replaced with a wide card to perform the search. [Selection diagram] Fig. 4