Voiceprint recognition method and device

Abstract

A method and system (800) for voiceprint recognition, include: establishing a first-level Deep Neural Network (DNN) model based on unlabeled speech data, the unlabeled speech data containing no speaker labels and the first-level DNN model specifying a plurality of basic voiceprint features for the unlabeled speech data; obtaining a plurality of high-level voiceprint features by tuning the first-level DNN model based on labeled speech data, the labeled speech data containing speech samples with respective speaker labels, and the tuning producing a second-level DNN model specifying the plurality of high-level voiceprint features; based on the second-level DNN model, registering a respective high-level voiceprint feature sequence for a user based on a registration speech sample received from the user; and performing speaker verification for the user based on the respective high-level voiceprint feature sequence registered for the user.

Description

technical field [0001] The embodiments of the present invention relate to the technical field of identity authentication, and more specifically, to a voiceprint recognition method and device. Background technique [0002] Voiceprint Recognition (VPR) is a kind of biometric technology, also known as speaker recognition (Speaker Recognition). Speaker identification includes two categories, namely speaker identification (Speaker Identification) and speaker confirmation (Speaker Verification). Speaker identification is used to judge which of several people said a certain speech, it is a "multiple choice" problem; and speaker confirmation is used to confirm whether a certain speech is spoken by a specified person, it is " One-to-one discrimination" problem. [0003] Voiceprint recognition includes two types: Text-Dependent and Text-Independent. The text-related voiceprint recognition system requires users to pronounce according to the specified content. Each person’s voiceprin...

AI Technical Summary

Problems solved by technology

In the application scenario of speaker recognition technology, the collected voice data is unlikely to be clean, and the types of noise contained in it are complex, and the signal-to-noise ratio is very poor.

If the traditional underlying spectrum-based features are used, after feature extraction, a large number of noise compensations at the feature end, model end, and score end are required. The calculation complexity is large, the time consumption is too long, and the influence of noise cannot be completely eliminated.

Images