DATA PROCESSING METHOD FOR OPTICAL CHARACTER Recognition (OCR), where the output data includes character images visually impaired

摘要

1. A method for resolving conflicting output from the optical character recognition system (OCR), wherein the output data includes at least one pattern visually impaired character, which is caused, at least one handling the object in the document It is processed by OCR system, wherein the method comprises: a) search for the output data for identifying images of characters having an image quality which exceeds a predetermined level, and using these character images s as a set of template images for characters; ! b) the location area identification portion of the visually impaired, at least one sample visually impaired character; ! c) use of the information on the location and area of ​​the visually impaired character instance to find the corresponding regions on the reference image before comparing the respective reference images for at least one character model visually impaired, neglecting the comparison image content to found their respective fields, and use a reference image having the greatest similarity to the image of the visually impaired character as a correct identifier ation visually impaired character. ! 2. A method according to claim 1, characterized in that the values ​​of the gray levels of pixels, concluded in the found region, hiding portion is assigned the value of the background level. ! 3. A method according to claim 1, characterized in that the symbol sample image of the visually impaired and the corresponding reference images correlator