Speech feature enhancement post-filtering method based on deep neural network

A deep neural network, speech feature technology, applied in speech analysis, instruments, etc., can solve problems such as non-stationary noise disadvantage, achieve the effect of suppressing noise interference and ensuring speech quality

Inactive Publication Date: 2020-06-26
CHONGQING UNIV OF POSTS & TELECOMM
View PDF2 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the estimation of the local minimum may have a delay of up to D frames (D is the analysis window length of the minimum statistical algorithm), which is very unfavorable for tracking non-stationary noise with rapidly changing power

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Speech feature enhancement post-filtering method based on deep neural network
  • Speech feature enhancement post-filtering method based on deep neural network
  • Speech feature enhancement post-filtering method based on deep neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0035] Embodiments of the present invention are described below through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation modes, and various modifications or changes can be made to the details in this specification based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the diagrams provided in the following embodiments are only schematically illustrating the basic concept of the present invention, and the following embodiments and the features in the embodiments can be combined with each other in the case of no conflict.

[0036] Wherein, the accompanying drawings are for illustrative purposes only, and represent only schematic diagrams, rather than physical drawings, and should...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a speech feature enhancement post-filtering method based on a deep neural network, and belongs to the technical field of speech filtering, and the method comprises the following steps: S1, mixing pure speech with noise according to different signal-to-noise ratios, and generating training data; S2, selecting the logarithm power spectrum LPS of the training data as a feature for extraction, and taking the LPS of the pure voice as a target; S3, performing training by using a standard structure deep neural network DNN; S4, performing loss estimation on a training result;and S5, inputting a loss voice, and performing loss compensation based on loss estimation. Compared with the prior art, noise interference can be effectively suppressed while the voice quality is ensured.

Description

technical field [0001] The invention belongs to the technical field of speech filtering, and relates to a speech feature enhancement post-filtering method based on a deep neural network. Background technique [0002] Speech is the most natural and commonly used way of information transmission for human beings. Whether in daily life or on the fast-growing Internet, voice, as one of the main media, carries a lot of useful information. Therefore, analyzing, processing and recognizing the information in the voice undoubtedly has broad application prospects. [0003] At present, there are two strategies to improve the noise robustness of the speech recognition system: ① reduce the influence of noise on the features to adapt to the acoustic model of clean speech training, that is, feature enhancement; ② modify the acoustic model to adapt to noisy speech , the model compensation. In contrast, feature enhancement runs at the front end of the recognition system and has a small tim...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G10L21/02G10L21/0208G10L25/21G10L25/30
CPCG10L21/0208G10L25/21G10L25/30
Inventor 武鹏飞周翊
Owner CHONGQING UNIV OF POSTS & TELECOMM
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap