Found a Perfect OCR method [PaddleOCR]
Suvoo opened this issue · comments
PaddleOCR - https://github.com/PaddlePaddle/PaddleOCR
Readme [English] - https://github.com/PaddlePaddle/PaddleOCR/blob/main/README_en.md
This OCR technique works better than Tesseract, EasyOCR and KerasOCR for low res images, small font texts and certain documents. I have tested this on certain images, should I make a PR on the readme adding this method?
It has an Apache License.