Title:
VOICEPRINT REGISTRATION METHOD BASED ON VOICEPRINT RECOGNITION MODEL, TERMINAL DEVICE, AND STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2019/134247
Kind Code:
A1
Abstract:
The present application discloses a voiceprint registration method based on a voiceprint recognition model, applied to a terminal device. The method comprises: acquiring valid voice of a user during voice registration; dividing the valid voice evenly into an integer number of parts of voice; calculating a characteristic voice vector of each part of voice among the integer number of parts of voice; determining whether each part of voice belongs to the same user according to the calculation result of the characteristic voice vector of each part of voice; and if said each part of voice belongs to the same user, registering the valid voice. The present application further provides a terminal device and a storage medium. The invention of the present application can detect whether a user is performing voiceprint registration in an environment where there is excessive noise or many people are speaking, and prevent voice recording not satisfying requirements from entering a voiceprint database (or a voice database), ensuring the quality of registered voice, preventing low-quality voice from affecting subsequent verification, and improving the usability of voiceprint recognition.
Inventors:
WANG JIANZONG (CN)
ZHENG SIQI (CN)
YU XIPAN (CN)
XIAO JING (CN)
ZHENG SIQI (CN)
YU XIPAN (CN)
XIAO JING (CN)
Application Number:
PCT/CN2018/077670
Publication Date:
July 11, 2019
Filing Date:
February 28, 2018
Export Citation:
Assignee:
PING AN TECH SHENZHEN CO LTD (CN)
International Classes:
G10L25/24
Foreign References:
| CN107342077A | 2017-11-10 | |||
| US20160248768A1 | 2016-08-25 | |||
| CN107146601A | 2017-09-08 | |||
| CN106601258A | 2017-04-26 |
Attorney, Agent or Firm:
SHENZHEN WORLD INTELLECTUAL PROPERTY AGENCY (GENERAL PARTNERSHIP ) (CN)
Download PDF: