OCR Systems

Image → Pre-process → Text Detection → Text Recognition → Post-process
        (denoise,    (where is text?)   (what does it say?)  (spell, format)
         deskew,
         binarize)

# Linux
sudo apt install tesseract-ocr tesseract-ocr-ben
pip install pytesseract

import cv2, pytesseract

img = cv2.imread("receipt.jpg")
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = cv2.threshold(gray, 0, 255,
       cv2.THRESH_BINARY + cv2.THRESH_OTSU)[1]

text = pytesseract.image_to_string(gray, lang="eng+ben")
print(text)

# word-level box
data = pytesseract.image_to_data(gray, output_type=pytesseract.Output.DICT)

# pip install easyocr
import easyocr

reader = easyocr.Reader(["en", "bn"], gpu=False)
results = reader.readtext("signboard.jpg")
for bbox, text, conf in results:
    print(f"{conf:.2f}  {text}")

pip install paddleocr paddlepaddle

from paddleocr import PaddleOCR
ocr = PaddleOCR(use_angle_cls=True, lang="en")
res = ocr.ocr("doc.jpg", cls=True)
for line in res[0]:
    box, (text, conf) = line
    print(conf, text)

import cv2, numpy as np

def deskew(gray):
    coords = np.column_stack(np.where(gray > 0))
    angle = cv2.minAreaRect(coords)[-1]
    angle = -(90 + angle) if angle < -45 else -angle
    (h, w) = gray.shape
    M = cv2.getRotationMatrix2D((w//2, h//2), angle, 1.0)
    return cv2.warpAffine(gray, M, (w, h),
                          flags=cv2.INTER_CUBIC,
                          borderMode=cv2.BORDER_REPLICATE)