Abstract: Image/text filtering apparatus and method for use in optical character recognition (OCR) scanning is disclosed. The invention filters video data representing text and image data on a document and erases the image data so that only the text data remains.

Abstract: A document scanner for utilization with a word processing and/or graphics processing equipment, such as a personal computer environment. The document scanner provides improved pixel compensation and gamma compensation techniques as well as improved control in the illumination process of the scanning of a document.

Abstract: Proportional spaced text recognition apparatus and method is disclosed. The invention is provided for optical character recognition (OCR) systems and provides recognition of both proportional spacing and fixed pitch type formats. The invention also provides recognition of accented characters, which are a common occurrence in Western European type texts.

Abstract: Low cost, high-speed, optical character isolation and page reconstruction system, method and apparatus are presented which overcome problems caused by copied pages, noise, underlines, skewed and bowed text, forms features, logos and signatures. As characters or noise are isolated and recognized, their corresponding bit patterns in memory are deleted. Recognized characters are isolated within entire words at a time to form page image records which are then linked to form lines of words. The text on the original page is then reconstructed from lines of words to yield output signals suitable for input to a host word processor.

Abstract: A character recognition system is disclosed utilizing a dead-band correlator for providing recognition of printed typestyles having horizontal and vertical stroke width variations without impairing the resolution required for character feature analysis. The system provides fewer character-to-mask registration errors, simultaneous computation of correlation scores of registration positions of masks with respect to the unknown character to compensate for additional registration errors, improved reject and substitution rates by utilizing unique threshold and separation requirements for masks, lower error rates by using small and large noise filtering and combining dual level acceptance criteria used in conjunction with re-try methods, stroke width normalization to aid in recognition of characters with badly degraded stroke widths, and selection of specific mask sets during multiple typestyle recognition processing than has previously been possible.

Abstract: An apparatus and method is described for the separation of data from adjacent characters of standard type fonts, some of which character pairs may kern or touch. Characters which do not kern or touch are separated by white column detection. Characters which do kern are first detected by a kerning test, which consists of locating white bits which separate the characters while meeting pre-established standards of contiguity. Touching characters are detected by failure to pass the white column test, followed by failure to pass the kerning test. Characters which touch are separated by a statistical analysis, which involves determination of which of several probable vertical data columns has the least number of character bits. Following separation, the characters are compared with pre-established character patterns.