Computer Recognition of Tamil, Malayalam and Devanagari Characters
Raghu Chandrasekaran, M. Chandrasekaran, Gift Siromoney · IETE Journal of Research · 1984
Computer recognition of multifont Tamil characters and special sets of printed Malayalam and Devanagari characters is described. Each character is converted manually into a rectangular binary array in which a zero represents a blank, and a one, a nonblank. Using the pattern encoding techniques developed by us, features in the form of strings are extracted by row-wise and column-wise scanning of the array. These features depend upon the number and nature of runs of ones in the rows and columns. Feature strings of printed prototype characters of an alphabet form a dictionary. Printed test input characters are fed into the computer, one by one, for recognition. Feature strings of each of the test input characters are obtained and compared with those of the prototype characters. A perfect match between the feature strings of the test input with a prototype would indicate that the two are identical and the test character is recognized as the corresponding prototype. Recognized characters are printed out in Roman letters following a special scheme of transliteration.