OCR (Optical Character Recognition) Software Development.
The project effort resulted in an evaluation of a correlation technique as implemented in an existing computer software program, as applied to automatic recognition of typeset printed characters which were sampled and digitized on a high-resolution scanner.
Abstract
Abstract : The project effort resulted in an evaluation of a correlation technique as implemented in an existing computer software program, as applied to automatic recognition of typeset printed characters which were sampled and digitized on a high-resolution scanner. Examples of Cyrillic and Latin text were extracted from several Russian technical journals. Also recorded were examples of non-character figures such as wood-chip and print-through blemishes, plus some graphics material. The data were edited and prepared and the results viewed on DICIFER, the RADC image-processing facility. The bulk of the correlation processing was performed using existing software (somewhat modified) from the DIMES package, and run on the RADC HIS 635 and 6180 computers.