Remote authorization method, system and device based on voiceprint recognition and storage medium

Abstract

The invention provides a remote authorization method, system and device based on voiceprint recognition and a storage medium. The method comprises the steps: receiving the voice of a user, and a customer service and the user conduct questioning and answering of a preset authentication question; cutting the received real-time audio of the user voice into multiple sections of sub-audios in real timeaccording to a preset duration as a unit; extracting audio features from the sub-audios, and inputting the audio features into a deep learning network to obtain voiceprint feature information of thesub-audios; performing cosine similarity comparison on the voiceprint feature information of each segment of sub-audio and voiceprint feature information pre-stored by the user to obtain a similarityvalue; judging whether the similarity average value meets a preset similarity threshold value or not, and if yes, determining that the authentication succeeds and the user obtains preset authorization; according to the invention, the order authorized by the authorizer can be directly processed, automation is realized, a lot of manpower is saved, identity authentication is completed in the oral authorization process of the authorizer, the security detection link is perfected, and the user benefit is maintained.

Description

technical field [0001] The invention relates to the field of remote authorization, in particular to a remote authorization method, system, device and storage medium based on voiceprint recognition. Background technique [0002] Voiceprint recognition is a kind of biometric technology, that is, speaker recognition, which can be divided into speaker recognition and speaker confirmation, text-related and text-independent. Speaker identification is to judge which of several speakers a certain voice belongs to, that is, a 1-to-many problem; speaker confirmation is to judge whether a certain audio is spoken by a specified speaker, that is, a 1-to-1 problem. Different scenarios use different voiceprint recognition technologies. Text-dependent requires the speaker to pronounce according to the specified content, while the text-independent recognition system does not limit the speaker's speech content. The former has a better recognition effect, while the latter has a wide range of...

AI Technical Summary

Problems solved by technology

In the case of oral authorization, there is an act of an unauthorized person using the identity of the authorizer to authorize the telephone. When the real authorizer is unknown, this behavior seriously damages the interests of both parties.

Images