How to identify single digits from image?

Question

I am trying to classify an image based on its content. For example, I have got loads of images as below, that will contain some content – in this case numeric values. I had tried OpenCV and Pytesseract OCR solution as proposed here: https://stackoverflow.com/a/60161328/7250310

However, this solution doesn't work on my images, and the content isn't detected. Below are my sample images:

Image 1:

Image 2:

Image 3:

Image 4:

Do you have any other ideas to achieve this? Basically Image 1 should give output as 1, and so on.

Answer 1

This simple approach works at least for the four presented images:

import cv2
import pytesseract

images = ['4sXGS.jpg', 'Nizki.jpg', 'T0EM8.jpg', 'g2fY7.jpg']

for img in images:

    img = cv2.imread(img, cv2.IMREAD_GRAYSCALE)
    img = cv2.threshold(img, 0, 255, cv2.THRESH_OTSU + cv2.THRESH_BINARY_INV)[1]

    text = pytesseract.image_to_string(img, config='--psm 10')
    text = text.replace('\n', '').replace('\f', '')
    print(text)

Output:

1
2
3
4

The single steps are:

1. Read the image as grayscale.
2. Inverse binary threshold the image using Otsu's method.
3. Run pytesseract using the -psm 10 option (single character). Maybe also add the described whitelisting for identifying digits only.

Caveat: I use a special version of Tesseract from the Mannheim University Library.

----------------------------------------
System information
----------------------------------------
Platform:      Windows-10-10.0.19041-SP0
Python:        3.9.1
PyCharm:       2021.1.1
OpenCV:        4.5.2
pytesseract:   5.0.0-alpha.20201127
----------------------------------------