Read image and speek the text content using PIL, langdetect, tesseract, gtts and playsound with Python
You must have to install tesseract-ocr on your local pc
- -i --image - path to input image to be OCR
- -p --processing - For type of pre-processing to be done(thresh, blur)
- -f --format - Format for the output(speak)
Read the Image
image = cv2.imread(args['image'])
Convert image to grayscale image
gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
Process the Image for threshold and blur according to user's choice For Threshold
gray_image = cv2.threshold(gray_image, 0, 255,
cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]
For Blurr
gray_image = cv2.medianBlur(gray_image, 3)
Store the image into temp folder with name as process_id of progrma
filename = "temp" + os.path.sep + "{}.png".format(os.getpid())
cv2.imwrite(filename, gray_image)
set path for tesseract
pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe'
get text from image
text = pytesseract.image_to_string(Image.open(filename))
remove temp file
os.remove(filename)
Store the processed image into output folder
text_filepath = "output" + os.path.sep + "{}.txt".format(os.getpid())
with open(file=text_filepath, mode='w+') as file:
file.write(text)
Detect the language accourding to the text
lang = detect(text)
print("Language: %s" % lang)
Convert Text to Audio using gtts
audio = gTTS(text=text, lang=lang, slow=False)
Store the audio file to output folder
audio_filepath = "output" + os.path.sep + "{}.mp3".format(os.getpid())
audio.save(audio_filepath)
write Grayimage to output folder
cv2.imwrite(filename, gray_image)
Play the converted audio
playsound(audio_filepath)
Sameer Narkhede
Profile : https://github.com/narkhedesam
Website : https://narkhedesam.github.io/
If this project help you reduce time to develop, you can give me a cup of coffee
