Identity authentication method and system

Abstract

The invention discloses an identity authentication method and an identity authentication system. The method comprises the following steps of: in the login of a user, receiving a continuous voice signal recorded by the current login user; extracting a voiceprint characteristic sequence from the continuous voice signal; computing likelihood between the voiceprint characteristic sequence and a background model; computing the likelihood between the voiceprint characteristic sequence and a speaker model of the current login user, wherein the speaker model is a polyhybrid Gaussian model constructed according to the repetition times and frame number of the registration voice signals recorded in the login of the current login user; computing a likelihood ratio according to the likelihood between the voiceprint characteristic sequence and the speaker model and the likelihood between the voiceprint characteristic sequence and the background model; and if the likelihood ratio is greater than a preset threshold value, determining the current login user is an effectively authenticated user, otherwise determining the current login user is an unauthenticated user. By the method and the system, the voiceprint-password-based identity authentication accuracy can be improved.

Description

technical field [0001] The invention relates to the technical field of identity identification, in particular to an identity authentication method and system. Background technique [0002] Voiceprint recognition (Voiceprint Recognition, VPR) is also called speaker recognition, and there are two types, namely, speaker identification and speaker confirmation. The former is used to judge which of several people said a certain speech, which is a "multiple choice" problem; while the latter is used to confirm whether a certain speech is spoken by a specified person, which is "one-to-one discrimination". "question. Different tasks and applications use different voiceprint recognition technologies. [0003] Voiceprint authentication refers to confirming the identity of the speaker based on the collected voice signal, which is a "one-to-one" discrimination problem. Today's mainstream voiceprint authentication system adopts a framework based on hypothesis testing, which is confirme...

AI Technical Summary

Problems solved by technology

The existence of redundant model parameters easily leads to an increase in storage and computing pressure, which in turn affects the efficiency of decoding

[0009] 2. The amount of model training is large: In the adaptive algorithm, it is necessary to calculate the sample statistics of each Gaussian of 1024 or even larger Gaussian numbers of the initial background model, and update its parameters

[0010] 3. In the adaptive algorithm, because the variance reestimation of the speaker model is difficult, the variance of the background model is often used directly

Directly using the variance of the background model cannot reflect the characteristics of the speaker model well, which reduces the distinction between different speaker models, thus affecting the recognition accuracy

Images