The main motive behind this project was that we often faced the problem of separately typing any content instead of copy-pasting from an already existing document or image which are not in typed format.
Hence, a text extractor which would simply scanning and extracting the content of the file would save loads of time and also reduce the chances of typographical error to 0%.
- Desktop Version (For images & PDF's) : http://18.222.220.89:5000/ (currently AWS instance is terminated)
- Mobile Version (Downloadable APK package only for images) : https://drive.google.com/file/d/1JLeMwO4LZjzQcbgoA5ARX-z3L8OsZ4-M/view?usp=sharing
Our system takes the scanned image/document from the user as an input.
Then perform some image pre-processing techniques, like scaling, binarization and noise removal.
Use Optical Character Recognition using Tesseract Engine and extract the text.
The link http://18.222.220.89:5000/ lands on this page, where you can submit the file from which you want to extract text.
After uploading and submitting a file, the result appear as shown in the image and you click on Copy To Clipboard to copy and use the text as you want.
After downloading the APK package from this link, install it in your device and start the app.
Upload an image and the results come out as follows. Then simply copy-paste the text and use it as per you requirement.
- Flask
- Tesseract OCR Engine
- TensorFlow, OpenCV
- Flutter
- AWS (Deployment)