Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
CORPUS SCREENING METHOD AND APPARATUS FOR SPEECH RECOGNITION TRAINING, AND COMPUTER DEVICE
Document Type and Number:
WIPO Patent Application WO/2020/224121
Kind Code:
A1
Abstract:
Provided are a corpus screening method and apparatus for speech recognition training, and a computer device and a computer-readable storage medium. The embodiments of the present application belong to the technical field of speech recognition. The method comprises: labeling corpora according to timestamps to obtain a first corpus set; training, by means of the first corpus set, a speech recognition model to obtain a first speech recognition model; decoding, by means of the first speech recognition model, each corpus segment in the first corpus set to obtain a first word sequence corresponding to each corpus segment; comparing each first word sequence with a standard word sequence corresponding to the first word sequence to calculate a first word recognition rate for each corpus segment; determining whether the first word recognition rate for each corpus segment meets a preset first word recognition rate condition; and storing the corpus segments corresponding to the first word recognition rates that meet the preset first word recognition rate condition, so as to form a screened second corpus set.

Inventors:
WANG TAO (CN)
Application Number:
PCT/CN2019/103470
Publication Date:
November 12, 2020
Filing Date:
August 30, 2019
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
PING AN TECH SHENZHEN CO LTD (CN)
International Classes:
G10L15/06; G06F16/635; G10L15/16
Foreign References:
CN108711421A2018-10-26
CN109388743A2019-02-26
CN105989081A2016-10-05
Attorney, Agent or Firm:
SHENZHEN TALENT PATENT SERVICE (CN)
Download PDF: