OCR for Tamil

[George Hart]

Here is a note from the recent INFITT conference (on computing in
Tamil) held in Foster City.  I would remark that of all the
Brahmi-derived alphabets, Tamil is by far the easiest for
implementing OCR, as it has no ligatures and lacks voiced and
aspirate consonants.  G. Hart

Two OCR softwares were to be discussed during Tamil Internet 2002.
"Ponvizhi" , Tamil OCR software of Prof. V. Krishnamoorthy. VK could
not attend the conference but his commercial software (cost US$150?)
was sold during the TI2002. Ponvizhi currently recognizes only pure
Tamil Texts. You may get more details or order a copy
from the author:   LearnFun Systems, 67/30 (4-c), II Main Road,
Gandhi Nagar, Adayar, Chennai -600 020. Tel: 446 1739
email: profvk at softhome.net

A demo version of this software is included along with several other
Tamil Softwares in the TI2002 Conference CD.  You may order a copy
of this CD (US$10) from INFIT secretariat <secretariat at infitt.org>

Prof. AG Ramakrishnan of IISc Bangalore presented details of a bilingual
(Tamil-Roman) OCR software "Mozhi Vallan" and a demonstration of
it at TI2002.