Problems reading digit '1' from images

[OjasMadaan]

Hi all, I'm currently trying to read characters from number plates - it usually works accurately. Still, for some images, it mistakenly reads the digit '1' as the letter 'L' and also makes minor mistakes elsewhere. Any suggestions to improve the accuracy? Basically I have a cropped image which only contains the license plate and the following code is run : reader = easyocr.Reader(['en'], gpu=False) result = reader.readtext(cropped_array) and the outputs look like :

[iblub1]

You could try to:

• finetune the model on your dataset
• use some heuristics if you have expert knowledge
• change some hyperparameters for the default recognizer https://github.com/JaidedAI/EasyOCR/blob/627503d0e8aa123ad294e46502b51ff18cd11bb3/easyocr/easyocr.py#L299