Image Number 6 for United States Patent #5272766.
An optical character recognition system is disclosed which is capable of recognizing out-of-specification characters and custom fonts, in addition to standard fonts such as E13B, OCRA, OCRB and Farington 7B type fonts recorded on documents such as bank drafts and checks. The standard fonts represent predefined characters, and the custom fonts and out-of-specification characters represent undefined characters. The system comprises an optical scanner, a microprocessor, a read-only-memory (ROM), a random-access-memory (RAM) with a stored program and predetermined template data, and a terminal with a keyboard and display unit. The scanner optically scans each document and produces a plurality of grey-scale pixel values which are stored in the RAM under control of the microprocessor. The microprocessor processes the stored pixel values, and effectively locates and segments each character on the document. In accordance with the predetermined template data, the microprocessor recognizes the segmented character if predefined, and rejects the segmented character if undefined. The rejected character is then displayed on the display unit for the user to see and identify. In response to identification provided (specified) by the user via the keyboard, the system develops a custom template or pattern representing (defining) the rejected character, and stores the custom template in memory for future use in recognizing the character.