Scientists from the Tomsk University of Control Systems and Radioelectronics have created a mathematical model that recognizes speech and outperforms its counterparts on several indicators.
Innovators point out that this model could be the basis of a comprehensive system of personal identification and facilitate the control of “smart” technology with the help of voice.
According to experts, a method of personal identification through speech recognition has been developed using digital security technology used to control equipment and security systems in remote banking services.
Scientists from the Tomsk University of Control Systems and Radioelectronics have created a new mathematical model of speech recognition that exceeds the accuracy of its peers and provides accurate speaker identification.
Anton Konev, Deputy Director of the National Communication Competence Center for “Reliable Interaction Technologies”, said: “The mathematical device differs from its peers because it recognizes the critical frequencies of speech and is very similar to the human inner ear .
The potential mechanisms underlying current device speech recognition systems are only helpful.”
Experts say that the error rate of the new device in recognizing the main tone of speech is four times lower than the best analogues in the world, which allows it to better recognize normal speech and even the singing. It also allows to communicate directly with smart devices, as with the interlocutor, not through commands.
Experts assure that the new system will allow verification of the identity of the interlocutor during the communication session, thereby completely eliminating some types of fraud.
Source: RIA Novosti
Source: Arabic RT
