Best practices for Tesseract OCR

[OlivierBinette]

1. Need to provide guidance for the use of Tesseract OCR:

   • What image resolution should be used?
   • What preprocessing should be applied for best performance?
   • What output format should be used?

2. Write function for OCR of newspaper scans using Tesseract and these best practices.

[OlivierBinette]

Tesseract's documentation has some guidelines for best practices: https://tesseract-ocr.github.io/tessdoc/ImproveQuality.html#rotation--deskewing

For us, the most important would be to deskew images, to remove borders, and to do some separation of the background. I'll follow up from #10 with visual examples of tesseract's output which showcase some issues we run into with the older newspaper scans.