Title:
A NEURAL-NETWORK-BASED APPROACH FOR SPEECH DENOISING STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH
Document Type and Number:
WIPO Patent Application WO/2022/107393
Kind Code:
A1
Abstract:
Disclosed are methods, systems, device, and other implementations, including a method that includes receiving an audio signal representation, detecting in the received audio signal representation, using a first learning model, one or more silent intervals with reduced foreground sound levels, determining based on the detected one or more silent intervals an estimated full noise profile corresponding to the audio signal representation, and generating with a second learning model, based on the received audio signal representation and on the determined estimated full noise profile, a resultant audio signal representation with a reduced noise level.
Inventors:
ZHENG CHANGXI (US)
XU RUILIN (US)
WU RUNDI (US)
VONDRICK CARL (US)
ISHIWAKA YUKO (JP)
XU RUILIN (US)
WU RUNDI (US)
VONDRICK CARL (US)
ISHIWAKA YUKO (JP)
Application Number:
PCT/JP2021/027243
Publication Date:
May 27, 2022
Filing Date:
July 20, 2021
Export Citation:
Assignee:
UNIV COLUMBIA (US)
SOFTBANK CORP (JP)
SOFTBANK CORP (JP)
International Classes:
G10L21/0208; G10L15/04; G10L15/16; G10L15/20; G10L21/0216; G10L21/0232; G10L21/0308
Foreign References:
| JPH06282297A | 1994-10-07 | |||
| JPH02253298A | 1990-10-12 |
Attorney, Agent or Firm:
RYUKA IP LAW FIRM (JP)
Download PDF: