Comment by Holybeds

There's a difference between doing OCR and actually understanding what is what in the document content.

For normal text OCR works well. But automatically understanding what is what is more complex.