MRZ extraction in passport using ppocr

[pbnewron]

Hi,
did anyone tried ppocr for MRZ extraction in passport? I was working with rec_v3 models and ocr performance is not very good compared to Tesserect. Does paddleocr out of box works for OCR-B font ( font used in MRZ section of passports)?
Can anyone suggest a good finetuning document or notebook, the one given in documentation dint help much?

regards
Prashant

[lishuoren]

Hello, I have encountered the same problem as you, did you solve the problem?

[Frosty-GTRIIP]

Hi there, I'm facing the same issue. It looks like the rec_v3 models occasionally fail to detect MRZ boxes. To address this, I'm currently retraining it as a detection model repeatedly. Have you found a solution to this problem?