Title:
VOICE ACTIVITY DETECTION METHOD AND APPARATUS, AND TERMINAL DEVICE AND COMPUTER STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2024/021270
Kind Code:
A1
Abstract:
Disclosed in the present invention are a voice activity detection method and apparatus, and a terminal device and a computer-readable storage medium. The method comprises: when a sound signal is collected by means of a microphone, performing feature extraction on the sound signal to obtain a feature sequence of the sound signal; inputting the feature sequence into a preset recursive neural network model for recurrent network training to obtain a model output signal, wherein the recursive neural network model is constructed on the basis of a recurrent neural network model having a gate structure; and confirming, according to the model output signal and a preset fully connected network, whether voice information is detected from the sound signal. By using the technical solution of the present invention, a model structure that is simple and requires fewer parameters can be constructed to perform voice activity detection, thereby effectively reducing the complexity of the whole voice activity detection process.
More Like This:
Inventors:
CHEN GUOMING (CN)
Application Number:
PCT/CN2022/120540
Publication Date:
February 01, 2024
Filing Date:
September 22, 2022
Export Citation:
Assignee:
GOERTEK TECH CO LTD (CN)
International Classes:
G10L25/78; G06N3/04
Foreign References:
CN113421595A | 2021-09-21 | |||
CN114333912A | 2022-04-12 | |||
US20200074997A1 | 2020-03-05 |
Download PDF: