Method for the recognition of script characters

摘要

Using a TV camera, the script image is registered and resolved line by line into individual pixels and stored. The individual script characters are analysed one after another. For this purpose, each script character field (1) is divided into four quadrants (2-5) and a lower extension (6). From each quadrant, vectors (11-16) are determined and, from these, a digital number is formed which is compared with a stored register. The method requires little storage capacity and computational expenditure and is therefore quick. Hence, the reliability can also be increased.