[tesseract-ocr] How to train tesseract with new script?

Hi all I am planning to train the ancient scripts for language translation. Is there any alternate rather than amazon mechanical turk to train the character? in stroke format. Or else have to train manually?? Thanks for taking time off your busy schedule... -- You received this message because

Re: [tesseract-ocr] How to train tesseract with new script?

oumik Ranjan Dasgupta < srd1...@cse.jgec.ac.in> wrote: > If you have a font of the said script alphabet, yes, I think it is > possible. > > On Thu, Apr 4, 2019, 11:01 PM Moni wrote: > >> Hi all >> I am planning to train the ancient scripts for language translation.

Re: [tesseract-ocr] How to train tesseract with new script?

ata . > > See > https://github.com/Shreeshrii/tessdata_coptic > > https://github.com/Shreeshrii/tessdata_shreetest/blob/master/brah.traineddata > > > > On Mon, Apr 8, 2019 at 9:13 AM Moni wrote: > >> Thanks for your valuable response >> Since the scripts doesn'

Re: [tesseract-ocr] How to train tesseract with new script?

er using real life images. > > I have tried training for Brahmi, Akkadian Cueniform and Coptic with > synthetic data . > > See > https://github.com/Shreeshrii/tessdata_coptic > > https://github.com/Shreeshrii/tessdata_shreetest/blob/master/brah.traineddata > > > >