Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
SPEECH PROCESSING METHOD, COMPUTER STORAGE MEDIUM, AND ELECTRONIC DEVICE
Document Type and Number:
WIPO Patent Application WO/2022/083039
Kind Code:
A1
Abstract:
A speech processing method and apparatus, a computer storage medium, and an electronic device, relating to the technical field of speech processing. The speech processing method comprises: acquiring a speech sequence, performing framing processing on the speech sequence to obtain a plurality of speech sub-sequences, and extracting target features of the speech sub-sequences (S110); detecting the speech sub-sequences by means of a speech detection model according to the target features, and determining valid speeches according to the detection results (S120); inputting the target features corresponding to the valid speeches into a voiceprint recognition model to screen a target speech from the valid speeches by means of the voiceprint recognition model (S130); and controlling to forward the target speech to other clients (S140). The method can filter and screen speech sequences, thus reducing a server bandwidth while improving speech transmission quality.

Inventors:
DING HANYU (CN)
LIN YUE (CN)
CHEN DUISHENG (CN)
Application Number:
PCT/CN2021/077309
Publication Date:
April 28, 2022
Filing Date:
February 22, 2021
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
NETEASE HANGZHOU NETWORK CO LTD (CN)
International Classes:
G10L21/0208; G10L17/02; G10L17/04; G10L17/06; G10L25/12; G10L25/18; G10L25/24; G10L25/30; G10L25/45
Foreign References:
CN112259114A2021-01-22
CN111028847A2020-04-17
CN108877775A2018-11-23
CN111755029A2020-10-09
CN111739515A2020-10-02
CN111599371A2020-08-28
CN111696580A2020-09-22
US20190392859A12019-12-26
Attorney, Agent or Firm:
BEIJING INTELLEGAL INTELLECTUAL PROPERTY AGENT LTD. (CN)
Download PDF: