Hlavní stránka Info portálu ZČU

All it takes is a short video in which a person's face is visible and a brief speech, such as introducing oneself, and the new technology from the laboratory of FAV UWB researchers will create a so-called hash, a sequence of binary coded numbers that are unique to that person. There is no need to store the person's original image or sound recording or their name anywhere, yet there is a unique key to identify them. The application uses artificial intelligence resources, precisely computer vision and speech recognition.
"The research aimed to prepare an application that would allow a person's identity to be easily verified, for example, instead of signing when receiving a parcel from a courier, so that it would be the least bothersome for that person and so that any data that could be considered personal or sensitive within the meaning of the European GDPR directive would not have to be stored during any stage of the verification process," explained Kamil Ekštein from FAV.

"If it is necessary to verify the identity of such a person, the smartphone app will again make an audiovisual recording of him or her so that he or she speaks either the same or a different speech as during the signature, and the software will evaluate whether the hash obtained from the verification recording matches the stored hash made during the signature," added Kamil Ekštein. The person is declared authenticated if the metrics defined within the technology match.

A significant advantage of Saynature technology, apart from the aforementioned GDPR compatibility, is the fact that the hash is relatively short and can be easily encoded, for example, into a QR code that can be printed and stuck on the parcel. The name of the Saynature technology is a play on words. It is a distortion of the word "signature" to indicate that when using this technology, you only have to say the signature.

The Saynature technology is now obtaining patent protection through U.S. Provisional Patent Application No. 63/601,118.