How to properly recognize the number in this kind of images?

Question

I'm trying to make a script that can identify the number in a picture, more precisely pictures REALLY similar to this one:

This goes from 50 to 1, but I'm having some problems reading the number present in there using pytesseract. Here's the code I'm using to read it:

im = Image.open(filename)
text = image_to_string(im)

All results I get are like this:

What can I do to improve the readings?

Answer 1

Improving the quality of the output is your "holy scripture" when working with Tesseract. Before binarization, you could first try to grayscale your image:

from PIL import Image
import pytesseract

im = Image.open('G9hvi.png').convert('L')
text = pytesseract.image_to_string(im)
print(text.replace('\f', ''))
# 50

Boom! – without any further pre-processing you already get the correct result.

----------------------------------------
System information
----------------------------------------
Platform:      Windows-10-10.0.19041-SP0
Python:        3.9.1
PyCharm:       2021.1.2
Pillow:        8.2.0
pytesseract:   5.0.0-alpha.20201127
----------------------------------------