OCR-Yolov5-SwinIR-SVTR

OCR(Korean)

Run demo

• Requirement

!python3 -m pip install paddlepaddle-gpu
!pip install "paddleocr>=2.0.1"
!pip install --upgrade PyYAML --ignore-installed
!pip install -r requirements.txt

• Run demo.py

!python3 demo.py --image_dir='/content/test_img/' --rec_algorithm='SVTR' --rec_image_shape='3,64,256' --rec_char_dict_path='./korean_dict.txt'

• Result : The result will be saved in './results'

Reference

• https://github.com/jingyunliang/swinir
• https://github.com/PaddlePaddle/PaddleOCR
• https://github.com/aqntks/Easy-Yolo-OCR