TungNT (Blue)

tungnt.blue@gmail.com

User Tools

Site Tools

Trace:
development:python:ocr

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
development:python:ocr [2024/08/30 03:24] – [Bounding Boxes with OpenCV] tungntdevelopment:python:ocr [2024/08/31 01:47] (current) – [Tessereact] tungnt
Line 46: Line 46:
 <code bash> <code bash>
 pip install pytesseract pip install pytesseract
 +</code>
 +
 +**Language:**
 +  * https://tesseract-ocr.github.io/tessdoc/Data-Files-in-different-versions.html
 +  * https://www.kaggle.com/code/dhorvay/pytesseract-multiple-languages
 +
 +<code python>
 +print(pytesseract.get_languages(config='.'))
 +
 +['afr', 'amh', 'ara', 'asm', 'aze', 'aze_cyrl', 'bel', 'ben', 'bod', 'bos', 'bre', 'bul', 'cat', 'ceb', 'ces', 'chi_sim', 'chi_sim_vert', 'chi_tra', 'chi_tra_vert', 'chr', 'cos', 'cym', 'dan', 'deu', 'div', 'dzo', 'ell', 'eng', 'enm', 'epo', 'equ', 'est', 'eus', 'fao', 'fas', 'fil', 'fin', 'fra', 'frk', 'frm', 'fry', 'gla', 'gle', 'glg', 'grc', 'guj', 'hat', 'heb', 'hin', 'hrv', 'hun', 'hye', 'iku', 'ind', 'isl', 'ita', 'ita_old', 'jav', 'jpn', 'jpn_vert', 'kan', 'kat', 'kat_old', 'kaz', 'khm', 'kir', 'kmr', 'kor', 'kor_vert', 'lao', 'lat', 'lav', 'lit', 'ltz', 'mal', 'mar', 'mkd', 'mlt', 'mon', 'mri', 'msa', 'mya', 'nep', 'nld', 'nor', 'oci', 'ori', 'osd', 'pan', 'pol', 'por', 'pus', 'que', 'ron', 'rus', 'san', 'sin', 'slk', 'slv', 'snd', 'snum', 'spa', 'spa_old', 'sqi', 'srp', 'srp_latn', 'sun', 'swa', 'swe', 'syr', 'tam', 'tat', 'tel', 'tgk', 'tha', 'tir', 'ton', 'tur', 'uig', 'ukr', 'urd', 'uzb', 'uzb_cyrl', 'vie', 'yid', 'yor']
 +
 +from langdetect import detect_langs 
 +
 +detect_langs(ocr_result_original)
 +
 +tesseract --list-langs
 </code> </code>
 ====== How to Open an Image in Python with PIL (Pillow) ====== ====== How to Open an Image in Python with PIL (Pillow) ======
Line 306: Line 322:
  
 ====== Introduction to PyTesseract ====== ====== Introduction to PyTesseract ======
 +
 +https://github.com/1sitevn/python-jupyter/blob/main/ocr/03_OCR_Introduction_To_PyTesseract.ipynb
  
 <file python> <file python>
Line 367: Line 385:
  
 ====== OCR an Index with PyTesseract ====== ====== OCR an Index with PyTesseract ======
 +
 +{{ :development:python:index_02.jpg |}}
 +
 +https://github.com/1sitevn/python-jupyter/blob/main/ocr/04_OCR_an_Index_with_PyTesseract.ipynb
  
 <file python> <file python>
Line 448: Line 470:
 print (results) print (results)
 </file> </file>
 +====== OCR a Text with Marginalia by Extracting the Body ======
 +
 +{{ :development:python:sample_mgh.jpg |}}
 +
 +https://github.com/1sitevn/python-jupyter/blob/main/ocr/06_OCR_a_Text_with_Marginalia_by_Extracting_the_Body.ipynb
 +
 +<file python>
 +import cv2
 +import pytesseract
 +from PIL import Image
 +
 +image = cv2.imread("../data/ocr/sample_mgh.JPG")
 +
 +base_image = image.copy()
 +
 +gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
 +blur = cv2.GaussianBlur(gray, (7,7), 0)
 +thresh = cv2.threshold(blur, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)[1]
 +
 +kernal = cv2.getStructuringElement(cv2.MORPH_RECT, (3, 50))
 +dilate = cv2.dilate(thresh, kernal, iterations=1)
 +
 +cv2.imwrite("../data/temp/sample_dilated.png", dilate)
 +
 +cnts = cv2.findContours(dilate, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
 +cnts = cnts[0] if len(cnts) == 2 else cnts[1]
 +cnts = sorted(cnts, key=lambda x: cv2.boundingRect(x)[1])
 +
 +for c in cnts:
 +    x,y,w,h = cv2.boundingRect(c)
 +    if h > 200 and w > 250:
 +        roi = base_image[y:y+h, x:x+w]
 +        cv2.rectangle(image, (x,y), (x+w, y+h), (36, 255, 12), 2)
 +
 +cv2.imwrite("../data/temp/sample_boxes.png", image)
 +
 +ocr_result_original = pytesseract.image_to_string(base_image)
 +
 +print(ocr_result_original)
 +
 +ocr_result_new = pytesseract.image_to_string(roi)
 +
 +print(ocr_result_new)
 +</file>
 +
 +====== Separate a Footnote from Body Text ======
 +
 +https://github.com/1sitevn/python-jupyter/blob/main/ocr/07_OCR_Separate_a_Footnote_from_Body_Text.ipynb
 +
 +<file python>
 +import cv2
 +import pytesseract
 +from PIL import Image
 +
 +image = cv2.imread('../data/ocr/sample_mgh_2.jpg')
 +base_image = image.copy()
 +
 +gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
 +blur = cv2.GaussianBlur(gray, (7,7), 0)
 +thresh = cv2.threshold(blur, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)[1]
 +
 +# Create rectangular structuring element and dilate
 +kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (3,25))
 +dilate = cv2.dilate(thresh, kernel, iterations=1)
 +
 +# Find contours and draw rectangle
 +cnts = cv2.findContours(dilate, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
 +cnts = cnts[0] if len(cnts) == 2 else cnts[1]
 +cnts = sorted(cnts, key=lambda x: cv2.boundingRect(x)[1])
 +main_text = ""
 +for c in cnts:
 +    x,y,w,h = cv2.boundingRect(c)
 +    if h > 200 and w > 250:
 +        roi = base_image[y:y+h, 0:x]
 +#         cv2.rectangle(image, (0, y), (x, 0 + h+20), (36,255,12), 2)
 +        
 +        constant= cv2.copyMakeBorder(roi.copy(),30,30,30,30,cv2.BORDER_CONSTANT,value=[255,255,255])
 +        ocr_result = pytesseract.image_to_string(constant)
 +        cv2.imwrite("../data/temp/output.png", roi)
 +        
 +        print (ocr_result)
 +#         print (ocr_result)
 +# cv2.imwrite("temp/output.png", image)
 +</file>
 +
 ====== References ====== ====== References ======
  
Line 454: Line 561:
   * https://www.affinda.com/tech-ai/how-to-convert-image-to-text-using-python   * https://www.affinda.com/tech-ai/how-to-convert-image-to-text-using-python
   * https://tesseract-ocr.github.io/tessdoc/Compiling.html#macos   * https://tesseract-ocr.github.io/tessdoc/Compiling.html#macos
 +  * https://nanonets.com/blog/ocr-with-tesseract/
development/python/ocr.1724988242.txt.gz · Last modified: 2024/08/30 03:24 by tungnt
Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki