Unlock instant, AI-driven research and patent intelligence for your innovation.

Speech Signal Sparse Representation Method Based on Convolutional Framework

A speech signal and sparse representation technology, applied in speech analysis, instruments, etc., can solve the problem of low data adaptability and achieve the effect of good data adaptability, good sparse representation ability, and robust numerical solution

Active Publication Date: 2021-05-14
NAT UNIV OF DEFENSE TECH
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Note that the local basis and non-local basis in the convolution framework are generally selected as basis matrices with structural characteristics, such as Fourier basis, cosine basis and wavelet basis. Although their structure is simple, the data adaptability is low

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 Signal Sparse Representation Method Based on Convolutional Framework
  • Speech Signal Sparse Representation Method Based on Convolutional Framework
  • Speech Signal Sparse Representation Method Based on Convolutional Framework

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0084] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described examples are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0085] The present invention provides a sparse representation method for speech signals based on a convolution framework. First, an optimization model of the convolution framework is established for a given non-local basis and a speech training signal set; figure 2 Shown) to realize the optimization training of the convolution framework, refer to figure 1 , figure 1 Flowchart for optimizing the training convolutional framework using the speech training signal set. F...

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

A method for sparse representation of speech signals based on convolutional framework. Aiming at the given non-local basis U in the convolutional framework, the speech training signal set is used to establish an optimized model of local basis V. After numerically solving the model, the non-local basis Based on the best matching local basis, the convolutional framework for optimal training is obtained. The convolution framework after optimization training can fully mine the local and non-local information of the speech signal to be represented, and has better sparse representation ability of the speech signal. The convolution framework after optimization training can utilize the structural features of the speech training signal set and has better data adaptability.

Description

technical field [0001] The invention relates to a speech signal sparse representation method, in particular to a speech signal sparse representation method based on a convolution framework. Background technique [0002] Sparse representation of speech signals has important applications in the fields of speech signal denoising, speech restoration, etc., which stems from a basic fact: natural signals have sparse characteristics, that is, when the signal is linearly decomposed on a representation base (dictionary), its representation coefficient is sparse. The Fourier base, cosine base, wavelet base, etc. in harmonic analysis provide important mathematical tools for the sparse representation of speech signals. Although its structure is simple and the amount of calculation is small, the sparse representation ability is limited, so the overcomplete base (dictionary) Came into being. The over-complete dictionary enhances the sparse representation ability, but the training is com...

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): G10L25/27
CPCG10L25/27
Inventor 王泽龙袁翰刘吉英叶钒余奇严奉霞朱炬波
Owner NAT UNIV OF DEFENSE TECH