synthetics_handwritten_OCR/OCR_earsing/utils/easy_ocr_utils.py

18 lines
388 B
Python
Raw Normal View History

import cv2
import easyocr
def easy_ocr_extraction(
img_path: str
) -> dict:
org_img_bgr = cv2.imread(img_path)
reader = easyocr.Reader(['fr', 'en'])
result = reader.readtext(org_img_bgr)
word_info = []
for (bbox, text, _) in result:
word_info.append(
[bbox[0][0], bbox[0][1], bbox[2][0], bbox[2][1], text]
)
return word_info