Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
METHOD AND SYSTEM FOR DETECTING VOICE ACTIVITY INNOISY CONDITIONS
Document Type and Number:
WIPO Patent Application WO/2020/043160
Kind Code:
A1
Abstract:
A voice activity detection method includes: training one or more computerized neural networks having a denoising autoencoder and a classifier, wherein the training is performed utilizing one or more models including Mel-frequency cepstral coefficients (MFCC) features, Δ features, ΔΔ features, and Pitch features, each model being recorded at one or more differing associated predetermined signal to noise ratios; recording a raw audio waveform and transmitting the raw audio waveform to the computerized neural network; denoising the raw audio wave utilizing the denoising autoencoder; anddetermining whether the raw audio waveform contains human speech; extracting any human speech from the raw audio waveform.

Inventors:
COSTELLO CHARLES (CN)
Application Number:
PCT/CN2019/103275
Publication Date:
March 05, 2020
Filing Date:
August 29, 2019
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
CLOUDMINDS SHENZHEN ROBOTICS SYSTEMS CO LTD (CN)
International Classes:
G10L15/04; G10L25/78
Foreign References:
CN101548313A2009-09-30
CN107851443A2018-03-27
CN108346428A2018-07-31
CN108346425A2018-07-31
US20130262111A12013-10-03
Attorney, Agent or Firm:
BEIJING ZHICHEN INTELLECTUAL PROPERTY AGENCY LTD. (CN)
Download PDF: