Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
DEEP LEARNING-BASED VOICEPRINT AUTHENTICATION METHOD AND DEVICE
Document Type and Number:
WIPO Patent Application WO/2017/201912
Kind Code:
A1
Abstract:
A deep learning-based voiceprint authentication method and device. The deep learning-based voiceprint authentication method comprises: receiving voice of a speaker (S11); extracting a d-vector feature of the voice (S12); obtaining a d-vector feature of the speaker determined at a registration stage (S13); calculating a matching value between the two d-vector features (S14); and determining that the speaker passes authentication if the matching value is greater than or equal to a threshold (S15). By means of the method, the voiceprint authentication effect can be improved.

Inventors:
WU, Bengu (3/F Baidu Campus, No.10 Shangdi 10th Street,Haidian District, Beijing 5, 100085, CN)
LI, Chao (3/F Baidu Campus, No.10 Shangdi 10th Street,Haidian District, Beijing 5, 100085, CN)
GUAN, Yong (3/F Baidu Campus, No.10 Shangdi 10th Street,Haidian District, Beijing 5, 100085, CN)
Application Number:
CN2016/098127
Publication Date:
November 30, 2017
Filing Date:
September 05, 2016
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
BAIDU ONLINE NETWORK TECHNOLOGY (BEIJING) CO., LTD. (3/F Baidu Campus, No.10 Shangdi 10th Street,Haidian District, Beijing 5, 100085, CN)
International Classes:
G10L17/04; G10L17/02
Domestic Patent References:
2015-10-15
Foreign References:
CN105869644A2016-08-17
CN104732978A2015-06-24
CN105575394A2016-05-11
Other References:
VARIANI, E. ET AL.: "Deep Neural Networks for Small Footprint Text-Dependent Speaker Verification", IEEE INTERNATIONAL CONFERENCE ON ACOUSTIC, SPEECH AND SIGNAL PROCESSING (ICASSP, 9 May 2014 (2014-05-09), XP032617560
LI, LANTIAN ET AL. ET AL.: "Improved Deep Speaker Feature Learning for Text-Dependent Speaker Recognition", PROCEEDINGS OF APSIPA ANNUAL SUMMIT AND CONFERENCE 2015, 19 December 2015 (2015-12-19), XP032870577
Attorney, Agent or Firm:
TSINGYIHUA INTELLECTUAL PROPERTY LLC (Room 301, Trade Building Zhaolanyuan,Tsinghua University, Qinghuayua, Haidian District Beijing 4, 100084, CN)
Download PDF: