Smart Engines secures US patent for on-device document recognition solution
An AI-driven solution developed by Smart Engines researchers which can be used to efficiently localize and identify documents in images has been awarded a patent in the United States.
The patent was registered on 7 February, bringing an end to a process that started many years ago, according to an announcement.
The solution can work on all devices like servers and smartphones and the document feature algorithm can be chosen depending on the device. Running the document recognition process directly on an end-users device could ease compliance with data protection regulations, and for applications in low-connectivity environments, the patent document states.
According to the company, the technology is already in use via its Smart ID Engine platform for recognition of ID cards, passports, and other ID credentials.
“This makes it possible to determine the type of document and eliminate distortions if necessary, i.e. to restore the original coordinate system. The invention differs from other works in that we analyze both local and global features on the image (from individual characters to the borders of the document, its colour, brightness, etc.),” explained Natalya Skoryukina, one of the inventors of the solution.
“The system is arranged in such a way that it is easy to add new types of documents and remove those that are no longer supported. It does not require retraining and a single prototype image is enough, where you don’t even need to see the parts with personal data – you can just ‘blur’ them.”
Commenting after the patent registration, Smart Engines CEO Vladimir Arlazarov said: “We are happy that already, our second invention has been registered as intellectual property in the USA.” He adds that the new solution is also crucial for his company’s business growth in the country.
Smart Engines had its ID document scanning and detection software integrated by One Vert last December.