Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A Robust Speech Feature Extraction Method Based on Nonlinear Power Transform Gammachirp Filter

A technology of speech features and extraction methods, applied in speech analysis, speech recognition, instruments, etc., can solve problems such as poor anti-noise performance, improve recognition accuracy, improve anti-noise robustness, improve anti-noise ability and robustness sticky effect

Active Publication Date: 2021-02-19
JIANGNAN UNIV
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Traditional speech feature extraction can have good results in quiet environments, but in complex noise environments, such algorithms generally have the problem of poor anti-noise performance

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
  • A Robust Speech Feature Extraction Method Based on Nonlinear Power Transform Gammachirp Filter
  • A Robust Speech Feature Extraction Method Based on Nonlinear Power Transform Gammachirp Filter
  • A Robust Speech Feature Extraction Method Based on Nonlinear Power Transform Gammachirp Filter

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0037] 1. Introduction to basic theory

[0038] 1. Gammachirp filter

[0039] The Gammachirp filter is a nonlinear filter that conforms to the auditory characteristics of the human ear, and its time domain expression is:

[0040] g c (t)=at n-1 exp(-2πbERB(f r )t)·exp(j2πf r t+jclnt+jφ)u(t)

[0041] In the formula, a is the amplitude, the filter order n and the parameter b are responsible for adjusting the distribution of the gamma function, according to the reference, here n and b take the values ​​of 4 and 1.109 respectively, and f r is the center frequency of the filter, φ is the initial phase, and generally takes φ=0. ERB(f r ) is the frequency f r The equivalent rectangular bandwidth of the time filter, its calculation formula is: ERB(f r )=24.7+0.108f r , where c is the chirp factor, and its value range is generally [-3,3], and c is used as the frequency modulation parameter of the Gammachirp filter to make it different from the Gammatone filter. When c=0, the ...

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 provides a robust speech feature extraction method based on a nonlinear power transformation Gammachirp filter, and mainly solves the problem of sharply degrading the performance of a speech recognition system under a noise environment. According to the method, a Gammachirp filter group conforming to the auditory characteristics of a cochlea is utilized, filters are treated and optimized through compression normalization, and after a response coefficient is obtained, the nonlinear characteristics of a signal are processed by simulating a human ear auditory model through a piecewise nonlinear power function transformation process. Moreover, the method combines the techniques of relative spectral RASTA filtering, mean-variance normalization and time series filtering to furtherimprove the anti-noise robustness of speech features. The method of the invention can improve the recognition rate of a speech recognition system under a noise environment, improve the anti-noise robustness of the system, and meet use requirements of daily safety fields such as smart home, in-vehicle system and various identity security authentications.

Description

technical field [0001] The invention belongs to the field of pattern recognition and speech processing, and relates to a robust speech recognition method in a real noise environment. Specifically, it is a robust speech feature extraction method based on nonlinear power transform Gammachirp filter, which can be used in daily life such as smart home, vehicle system, etc., as well as in various security fields that require security authentication. Background technique [0002] Speech recognition-related systems, at their most basic level, are a collection of different approaches drawn from research in various fields and disciplines, including signal processing, pattern recognition, and speech linguistics. Each of these disciplinary approaches transforms the speech signal waveform into some type of parametric representation, which is known as speech feature extraction. Feature extraction is the most basic and important process in speech signal processing. A good feature extract...

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 Patents(China)
IPC IPC(8): G10L15/22G10L25/24G10L15/20G10L19/26G10L21/0208G10L25/27
CPCG10L15/20G10L15/22G10L19/26G10L21/0208G10L25/24G10L25/27
Inventor 葛洪伟李聪陈国俊
Owner JIANGNAN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com