Re: [tesseract-ocr] Training Tesseract for new fonts

2022-10-06 Thread Umanda Dikwatta
Thank you very much for the link. Can we use non-unicode fonts as well? I have attached a sinhala font that I'm struggling to train. Thank you very much On Thu, Oct 6, 2022 at 11:10 AM Saman Kurdi wrote: > Hello, > > This might help. > > https://www.mdpi.com/2076-3417/11/20/9752 > > Refards. >

Re: [tesseract-ocr] Training Tesseract for new fonts

2022-10-05 Thread Saman Kurdi
Hello, This might help. https://www.mdpi.com/2076-3417/11/20/9752 Refards. On Thu, Oct 6, 2022 at 07:37 Umanda Dikwatta wrote: > Hello, > > I've been using Tesseract 4.1 for some time. I am using Tesseract with > Sinhala language. I got good results for most of the images I tried. I > trained

[tesseract-ocr] Training Tesseract for new fonts

2022-10-05 Thread Umanda Dikwatta
Hello, I've been using Tesseract 4.1 for some time. I am using Tesseract with Sinhala language. I got good results for most of the images I tried. I trained Tesseract with different fonts. But as the documentation says, I had to preprocess my images to obtain good results. Then I tried Tesser