I want to detect blocks of text from the images. I want the model to be able to detect each of the blocks as paragraph or address or signature or table or title of the document.
I would want to use machine learning/deep learning using python and open source libraries available.
I am a novice at machine learning. Please help me out with this. If possible please share sample code for better understanding.
Thank you
Hi,
I would suggest the Book "Learning TensorFlow" by O'REILLY, which is in my opinion a very good starting book on the topic.
Furthermore there are code examples for - not exactly - problems like yours. You can learn to automatically detect handwritten numbers and process images. I think combining them could answer your question.
Additionally, if you never worked with machine learning until now, the book provides solid basics for you to understand everything.
Hope this helps you!
In my opinion refer this link getting some ideas
https://www.youtube.com/watch?v=TWJXCJA_hzI
I need to remember how I did _that_ about two months ago... and it was for scanned English texts OCR with English Lexicon... I wonder if there is any other, say, Russian / Chinese Lexicon for such an attempt... will update when I find more time. - s.ye
Varun Sk , you may try Pytesseract, its quite simple for beginners in machine learning yet its powerful enough to do extraction text and image from scanned documents.
PyTesseract is one of the best and easy way to extract text and images from scanned document images.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing