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

Silent voice decoding method based on surface electromyogram signals

A technology of myoelectric signal and speech decoding, which is applied in the field of accurate and natural silent speech recognition and silent speech decoding, can solve the problems that are difficult to meet the needs of actual speech communication, and achieve excellent phrase recognition accuracy and low decoding word errors rate and reduce the effect of word error rate

Pending Publication Date: 2022-08-05
UNIV OF SCI & TECH OF CHINA
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, the above-mentioned silent speech recognition method based on isolated words is difficult to meet the needs of actual speech communication.

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
  • Silent voice decoding method based on surface electromyogram signals
  • Silent voice decoding method based on surface electromyogram signals
  • Silent voice decoding method based on surface electromyogram signals

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0075] In this embodiment, a silent speech decoding method based on surface EMG signals, such as figure 1 shown, it includes the following steps:

[0076] Step 1. Construct a silent speech surface EMG dataset D:

[0077] Construct a corpus containing R Chinese phrases E={e 1 ,…,e v ,…,e R }, where e v Represents the vth Chinese phrase in the corpus, all R Chinese phrases are generated by a dictionary containing L different syllables; the surface EMG data of the user's silent pronunciation corpus are collected by flexible high-density electrode arrays and multi-channel signal conversion equipment, Then, the surface EMG signal data is divided into non-overlapping frames, and the time domain features of each frame are extracted respectively, so as to obtain M EMG signal feature samples, which are recorded as the data set D={(x 1 ,y 1 ),(x 2 ,y 2 ),...,(x i ,y i ),...,(x M ,y M )}, where x i represents the i-th EMG signal feature sample, and Represents the i-th EM...

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 discloses a silent voice decoding method based on surface electromyogram signals. The silent voice decoding method comprises the following steps: 1, collecting silent voice electromyogram data and extracting features to construct a training data set; 2, a decoding network based on a Transform model is built; and 3, training the decoding network, and obtaining an optimized and updated decoding network for realizing silent speech recognition based on the surface electromyogram signals. According to the method, the syllable sequence can be decoded from the silent voice electromyographic signal, so that the semantic information can be understood, and high-precision silent voice recognition can be realized.

Description

technical field [0001] The invention relates to the fields of electromyography signal processing and natural language processing, in particular to a silent speech decoding method based on surface electromyography signals, which is mainly applied to accurate and natural silent speech recognition. Background technique [0002] Speech can carry and transmit certain meaningful semantic information, and it is a natural and effective way of communication for human beings. In recent years, with the rapid development of computer science and technology, automatic speech recognition system has attracted much attention as an attractive human-computer interaction method. Automatic speech recognition technology can realize the conversion of speech signal to text information through computer. By integrating speech recognition interfaces into smart devices such as mobile phones, speakers, and cars, automatic speech recognition technology is widely used in people's daily lives. However, t...

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
IPC IPC(8): G10L15/02G10L19/16G10L25/66G06K9/00G06K9/62G06N3/04G06N3/08
CPCG10L15/02G10L19/167G10L25/66G06N3/08G10L2015/027G06N3/047G06N3/045G06F2218/08G06F2218/12G06F18/2415G06F18/214
Inventor 张旭宋芮陈希陈香陈勋
Owner UNIV OF SCI & TECH OF CHINA