On Mon, Oct 15, 2012 at 6:17 AM, Kishore Babu <kb...@envistacorp.com> wrote:
> Hi Peter, > Thank you very much for the reply. Unfortunately, the image I am dealing > are the scanned one. > > I will update my result if I succeed in using the mentioned line detection > algorithms. > > There is an excellent explanation of everything involved in OCR at : http://sudokugrab.blogspot.com.au/2009/07/how-does-it-all-work.html . This is a hard problem but the better your scan the easier it becomes. if you have good contrast, modern typefaces, no/little skewing, and a well understood character set then you have a chance. -- Peter Murray-Rust Reader in Molecular Informatics Unilever Centre, Dep. Of Chemistry University of Cambridge CB2 1EW, UK +44-1223-763069
