Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
LEARNING DEVICE, VOICE ACTIVITY DETECTOR, AND METHOD FOR DETECTING VOICE ACTIVITY
Document Type and Number:
WIPO Patent Application WO/2019/162990
Kind Code:
A1
Abstract:
The present invention corrects a level of sound serving as a discriminant measure to determine the boundary between noise and voice using a Gaussian mixture model for noise learned during a time interval when an input signal is noise, and detects a voice activity on the basis of the corrected level of sound.

Inventors:
HANAZAWA TOSHIYUKI (JP)
Application Number:
JP2018/005938
Publication Date:
August 29, 2019
Filing Date:
February 20, 2018
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
MITSUBISHI ELECTRIC CORP (JP)
International Classes:
G10L25/78
Foreign References:
Other References:
AKIHIRO: "investigation of acttivity deep neutral network and cross adaptation for voice activity detection in meeting speech", IEICE, vol. 114, no. 365, 8 December 2014 (2014-12-08), pages 19 - 24
TAMURA ET AL: "multi-modal speech recognition using deep bottleneck features", IEICE, vol. 115, no. 253, October 2015 (2015-10-01), pages 57 - 62
SEIJI, DEEP NEURAL NETWORK, March 2016 (2016-03-01), pages 6 - 23, Retrieved from the Internet [retrieved on 20180502]
Attorney, Agent or Firm:
TAZAWA, Hideaki et al. (JP)
Download PDF: