Tesseract training

[karim abed el hadi]

Hello everyone, 

My name is Karim Abed El Hadi, I am a telecommunications engineering student at the Holy Spirit University of Kaslik, Lebanon.

I am working on a Syriac OCR project.

For the first part of my project, I did a GUI that lets the user select an image or a PDF file from the device, and use it as an input to the OCR engine and it is working really well.

For the second part, I want to train the tesseract to improve the accuracy of the results. I am working with Python 3.8 and tesseract version 4.

1- What are the steps to train the tesseract v4?

2- Is there a way to train the tesseract automatically every time an input image is introduced to the tesseract (fine tuning)?

Karim Abed El Hadi