I'm doing text recognition in scanned text pages and recently started trying Tesseract. I realize it sometimes struggles with some tasks so I created a region of interest in a field where I will have none to two characters to recognize, like so:
I did some cleanup to get this "C" character, however, running the picture on Tesseract:
$ tesseract image.png stdout
It still fails to recognize it and I get the following output:
Warning. Invalid resolution 0 dpi. Using 70 instead.
Estimating resolution as 275
Empty page!!
Estimating resolution as 275
Empty page!!
Is there any reason for it to be failing at this seemingly simple task?