Voice discrimination method, device, electronic device and storage medium

Abstract

The present disclosure relates to a voice discrimination method, device, electronic equipment, and storage medium. The method includes: acquiring the voice to be differentiated; extracting the acoustic features and language style features of the voice to be differentiated; analyzing the acoustic features and the language style The features are subjected to speech splicing processing to obtain fusion features; the fusion features are input into a speech discriminator to determine whether the speech to be discriminated is real speech or generated speech. The above-mentioned technical means are adopted to solve the problems in the prior art that the accuracy rate of distinguishing generated speech and real speech is low, and the method for distinguishing generated speech and real speech does not have universal applicability.

Description

technical field [0001] The present disclosure relates to the field of voice recognition, in particular to a voice discrimination method, device, electronic equipment and storage medium. Background technique [0002] With the development of deep learning technology, the similarity between the generated speech obtained by using speech synthesis and speech conversion technology and the real speech of a real person has been greatly improved, and even the real one is false. It has a wide range of applications in medical, entertainment and other fields. It provides the technical conditions for using generated voice to carry out network fraud, which poses a great threat to people's safety and social stability. Therefore, the identification technology of generated voice has become an urgent need in today's society. In the prior art, acoustic features are often used to identify generated speech and real speech, but only acoustic features are used, and the generated speech and real sp...

AI Technical Summary

Problems solved by technology

[0002] With the development of deep learning technology, the similarity between the generated speech obtained by using speech synthesis and speech conversion technology and the real speech of a real person has been greatly improved, and even the fake one has been widely used in medical, entertainment and other fields. It provides the technical conditions for using generated voice to carry out network fraud, which poses a great threat to people's safety and social stability. Therefore, the identification technology of generated voice has become an urgent need in today's society

The existing technology often uses acoustic features to identify generated speech and real speech, but only using acoustic features cannot distinguish generated speech and real speech well

[0003] In the process of realizing the concept of the present disclosure, the inventors found that there are at least the following technical problems in the related art: the accuracy rate of distinguishing generated speech and real speech is low, and the method of distinguishing generated speech and real speech is not universal

Images