Reputation: 231
I am trying to recognize numbers in the attached image. As you can see, some parts of the numbers are cut off, making it challenging to achieve accurate recognition.
I attempted to fill in the missing parts of the numbers and make the image clearer, but I encountered issues such as the parts of the number "5" getting filled incorrectly. Do you have any suggestions or methods to improve the recognition accuracy?
I have tried using pytesseract, but it failed. EasyOCR had a slightly higher success rate, but the results are still not satisfactory. Here is the code I used:
import easyocr
reader = easyocr.Reader(['en'], gpu=True)
resu = reader.readtext('img.jpg', allowlist='0123456789')
Upvotes: 0
Views: 95