ARTICLE AD BOX
I write a tool (C# 9) to solve word games like crosswords and others. I am provided with the game board as an image.
I tried various ways to read the board in order to build an in-memory object representation of it.
I am not satisfied with my results for now.
How would you do that ?
using a 3rd party lib to extract text, such as Tesseract ?
preparing images of individual letters, then compare them with the board image ?
other ?
I aim precision more than performance.
This is an example of a board as image :
