Document processing
Despite the fact that humanity has been working towards reducing the amount of paper used for many years, still a huge number of documents are available only in paper form. Their processing is an important task that deep learning helps to solve.
Optical Character Recognition (OCR)
Optical text recognition allows you to translate text from paper documents into electronic documents.
Receipts data collection
Digitizing receipts is a necessary step for bookkeeping and expense tracking. Receipts are printed on thermal paper, so they tend to lose the clarity of the printed text over time and can become completely unreadable when overheated. Computer vision will help in cases where standard tools can no longer provide the necessary accuracy.
Writer identification
The analysis of handwritten text is one of the most difficult tasks in document processing, each person's handwriting is unique, which is a problem in text recognition, but at the same time allows one to identify the author. Deep learning allows you to do this analysis without a professional expert graphologist.
Signature verification
As a further development of the Writer identification task, there is the task of confirming that the signature on certain documents corresponds to this particular person. In the past, such problems were solved with the help of expensive experts, now artificial intelligence allows you to solve this problem quickly with the help of a few clicks.
Diagrams translation
Digitization of diagrams allows you to transfer printed and handwritten diagrams into digital form for further editing. The architecture of the application is drawn during brainstorms, the explanation on the whiteboard or the photo is taken from the official documentation - digitize it!
