CN108805067A - Surface electromyogram signal gesture identification method - Google Patents
Surface electromyogram signal gesture identification method Download PDFInfo
- Publication number
- CN108805067A CN108805067A CN201810552616.8A CN201810552616A CN108805067A CN 108805067 A CN108805067 A CN 108805067A CN 201810552616 A CN201810552616 A CN 201810552616A CN 108805067 A CN108805067 A CN 108805067A
- Authority
- CN
- China
- Prior art keywords
- signal
- gesture
- function
- intrinsic mode
- surface electromyogram
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2134—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on separation criteria, e.g. independent component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Signal Processing (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
A kind of surface electromyogram signal gesture identification method carries out variation mode decomposition to surface myoelectric hand signal first, obtains adaptive decomposition signal, and FastICA methods eliminate the noise in signal;Then method is locally linear embedding into the high dimensional data dimensionality reduction of hand signal and extracts feature using supervision, obtain the effective low-dimensional eigenmatrix of hand signal;Finally hand signal is classified and identified using support vector machine classifier.The present invention can carry out adaptive decomposition to surface electromyogram signal, eliminate the noise in signal, have stronger anti-interference ability and noise robustness;The present invention can recognize that different gestures and has higher Classification and Identification rate, the classification that can be used in terms of surface electromyogram signal gesture identification and other signals.
Description
Technical field
The invention belongs to actuating signal identification technology fields, and in particular to arrive the identification of various gestures.
Background technology
With the development of cross discipline research, surface electromyogram signal gesture identification is widely used in biomedical, rehabilitation
Engineering and artificial intelligence etc., this research have evolved into popular research in bio signal processing and area of pattern recognition
Project.Gesture identification can allow physical disabilities patient to efficiently accomplish limb action by artificial limb, be hopeful to return to the life of normal person
State living;The manual communication of deaf-mute and Healthy People can be made to become simple, it is easier to understand that mutual exchange is intended to;It can also
Keep people's lives more convenient and smart in somatic sensation television game and smart machine, improves the quality of living.
Surface myoelectric hand signal itself there are non-linear, non-stationary, randomness is strong, signal is weak the features such as, gatherer process
In ambient noise can be added.The maximum challenge of research surface myoelectric gesture identification is how processing with high dimension, non-linear
The data of feature, common feature extracting method are difficult the real structure obtained in hand signal data so that pattern-recognition becomes
It obtains difficult.
Single channel EMG Signal Decomposition Based is one group of natural mode of vibration letter using overall experience mode decomposition method by NaikG et al.
Number, using quick independent analytical methods burbling noise, 5 time domain spies are extracted using linear discriminant analysis method in separation component
Sign carries out simplifying classification, and this method has higher recognition capability to normal person and muscle disease patient.Zou Xiaoyang, Lei Min et al.
Angle non-linear from surface electromyogram signal, non-stationary proposes multiple dimensioned fuzzy entropy feature extracting method.It is sharp first
Multi-resolution decomposition is carried out to signal with wavelet decomposition, signal characteristic is extracted with multiple dimensioned fuzzy entropy, uses support vector machines pair six
Kind gesture motion carries out Classification and Identification.Zhang Qizhong, Xi Xugang et al. propose a kind of calculating Nonlinear Time Series signal
The new methods of Lyapunov indexes extracts signal characteristic, using support vector machines to clenching fist, exrending boxing, wrist inward turning, 4 class of wrist outward turning it is dynamic
Operation mode has carried out preferable Classification and Identification.
There are end effect, spectral aliasings when above-mentioned overall experience mode decomposition method decomposes surface electromyogram signal
Etc. technical problems, the application condition of signal decomposition it is big.The time domain of generally use in surface electromyogram signal research process, frequency domain, when
Frequency domain and the noiseproof feature of the analysis methods such as non-linear are weak, high dimensional data can be caused to occur so that the Classification and Identification of pattern-recognition
Rate is not high.
Invention content
Technical problem underlying to be solved by this invention is to overcome the shortcomings of above-mentioned prior art, provide a kind of anti-interference
Property strong, surface electromyogram signal gesture identification method that Classification and Identification rate is high.
Technical solution is made of following step used by solving above-mentioned technical problem:
(1) multistage intrinsic mode function is determined
Surface electromyogram signal gesture data collection is extracted from UCI databases, and table of arbitrary extraction is concentrated from gesture data
Facial muscle electric signal carries out variation mode decomposition, obtains the intrinsic mode function of formula (1):
α is penalty factor in formula,It is the Fourier transformation of gesture surface electromyogram signal,It is that Lagrange multiplies
The Fourier transformation of son,It is the Fourier transformation of i-th of intrinsic mode function,It is k-th natural mode of vibration letter
Several Fourier transformations, n are the newer numbers of expression formula (1), are the numeric field frequency that 0~4, ω is gesture surface electromyogram signal
Rate, ωk={ ω1,ω2, ,ωkBe whole intrinsic mode functions centre frequency, k is recycled to K, the mode that k is to determine from 1
It is 1~5 to decompose number;To intrinsic mode function progress and mode decomposition number phase homogeneous iteration, multistage natural mode of vibration letter is obtained
Number.
(2) denoising is carried out to multistage intrinsic mode function
Denoising is carried out using existing Fast Independent Component Analysis method using multistage intrinsic mode function as hand signal
Processing.
(3) feature vector of hand signal is obtained
Method is locally linear embedding into using supervision, dimensionality reduction is carried out according to a conventional method to the hand signal after denoising, extraction is special
Sign, obtains feature vector shown in formula (2):
It is limited positive integer that l, which is the sample number of each gesture, in formula, and i, j are limited positive integer, yiIt is embedded in for low-dimensional
Vector, yjIt is yiNeighborhood point, wijIt is weight coefficient.
(4) gesture identification
The feature vector input support vector machine classifier of all surface electromyography signal is concentrated to carry out gesture gesture data
Identification, obtains gesture identification result.
It is of the invention arbitrarily to be carried from gesture data concentration in the multistage intrinsic mode function step (1) of determination of the present invention
Taking one-time surface electromyography signal to carry out variation mode decomposition, steps are as follows:
(1) to intrinsic mode function uk(t) Hilbert transform is carried out, unilateral frequency spectrum is obtained:
δ (t) is unit impulse function in formula, and * is convolution, and j is imaginary unit, and t is time, uk(t) it is natural mode of vibration letter
Number, k is the number of intrinsic mode function.
(2) the centre frequency ω that index adjusts each modal components is added in each intrinsic mode functionk:
* is convolution in formula, and j is imaginary unit, and t is the time, and δ (t) is unit impulse function.
In the multistage intrinsic mode function step (1) of determination of the present invention, mode number K most preferably 4 of the invention is being obtained
It takes in the feature vector step (3) of hand signal, the sample l most preferably 150 of each gesture of the invention.
The present invention is compared with existing surface electromyogram signal gesture identification method with following features:
The present invention proposes variation mode decomposition method, avoids the endpoint effect occurred in overall experience mode decomposition method
Answer, spectral aliasing the problems such as;Variation mode decomposition method is used, the effective characteristic frequency of signal is obtained, can more disclose gesture
The rule of signal;Using Fast Independent Component Analysis method, can quickly noise be removed and be remained with from surface electromyogram signal
Surface myoelectric information;It uses supervision and is locally linear embedding into method, dimensionality reduction is carried out to surface electromyogram signal and extract feature,
It is excavated out from the nonlinear surface electromyography signal of higher-dimension complexity comprising multiple manifold geometric attribute, brief initial data simultaneously retains
The essential structure of data, saves the feature of signal.
Description of the drawings
Fig. 1 is the flow chart of the embodiment of the present invention 1.
Fig. 2 is the image of gesture form of the present invention.
Specific implementation mode
The present invention is described in more detail with reference to the accompanying drawings and examples, but the present invention is not limited to following embodiment party
Formula.
Embodiment 1
By for extracting 6 kinds of gesture data collection of surface electromyogram signal in UCI databases, acquire 5 ages 20~
The gesture surface electromyogram signal of 22 years old health volunteers, wherein male 2, women 3.It acquires subject and captures article completion
Holding cylinder (Fig. 2 a), portable weight (Fig. 2 b), hand-held flat articles (Fig. 2 c), palm when held object (Fig. 2 d), hold
Sphere (Fig. 2 e), hand pinch the surface electromyogram signal of six kinds of gestures of small article (Fig. 2 f).Each gesture repeats 30 times, everyone adopts
The hand signal of collection is stored with a MAT file, and there are five MAT files altogether, share 900 gesture surface electromyogram signals, including
150 holding cylinders, 150 hand-held flat articles, 150 palm when held objects, are held the puck for 150 times at 150 portable weights
Body, 150 hands pinch small article hand signal.Surface electromyogram signal gesture identification method is as shown in Figure 1, steps are as follows:
(1) multistage intrinsic mode function is determined
Surface electromyogram signal gesture data collection is extracted from UCI databases, and table of arbitrary extraction is concentrated from gesture data
Facial muscle electric signal carries out variation mode decomposition, obtains the intrinsic mode function of formula (1):
α is secondary penalty factor in formula,It is the Fourier transformation of gesture surface electromyogram signal,It is that glug is bright
The Fourier transformation of day multiplier,It is the Fourier transformation of i-th of intrinsic mode function,It is k-th natural mode
The Fourier transformation of state function, K 5, n are the newer numbers of expression formula (1), are the number that 4, ω is gesture surface electromyogram signal
Domain frequency, ωk={ ω1,ω2, ,ωkBe whole intrinsic mode functions centre frequency, k is recycled to K from 1, what K was to determine
Mode decomposition number is 1~5;To intrinsic mode function progress and mode decomposition number phase homogeneous iteration, multistage natural mode is obtained
State function.
It is above-mentioned to concentrate arbitrary extraction one-time surface electromyography signal to carry out variation mode decomposition method and step such as gesture data
Under:
1) to intrinsic mode function uk(t) Hilbert transform is carried out, unilateral frequency spectrum is obtained:
δ (t) is unit impulse function in formula, and * is convolution, and j is imaginary unit, t be 6 seconds the time, uk(t) it is natural mode
State function, k are the numbers of intrinsic mode function, are 1~5.
2) the centre frequency ω that index adjusts each modal components is added in each intrinsic mode functionk;
* is convolution in formula, and j is imaginary unit, and t is the time, and δ (t) is unit impulse function.
(2) denoising is carried out to multistage intrinsic mode function
Denoising is carried out using existing Fast Independent Component Analysis method using multistage intrinsic mode function as hand signal
Processing, the specific method of denoising exist《FastICA algorithm answering in type local-discharge ultrasonic array signal denoising
With》(Proceedings of the CSEE the 18th phase of volume 32 in 2012:160-166) Section 1 discloses.
(3) feature vector of hand signal is obtained
Method is locally linear embedding into using supervision, dimensionality reduction, supervision part are carried out according to a conventional method to the hand signal after denoising
Linearly embedding method carries out dimensionality reduction and exists《Fault Diagnosis of Aeroengines diagnostic method based on supervision manifold learning》(promote skill
Art the 5th phase of volume 38 in 2017:1147-1154) Section 2 discloses, and extracts feature, obtains feature vector shown in formula (2):
It is limited positive integer that l, which is the sample number of each gesture, in formula, and i, j are limited positive integer, yiIt is embedded in for low-dimensional
Vector, yjIt is yiNeighborhood point, wijIt is weight coefficient.
(4) gesture identification
The feature vector input support vector machine classifier of all surface electromyography signal is concentrated to carry out gesture gesture data
Identification, obtains gesture identification as a result, being shown in Table 1.
The Classification and Identification rate (%) of the different training samples of table 1
By table 1 as it can be seen that when number of training and test sample number are all respectively 75 times, the average classification of six kinds of gestures is known
Rate is not 93.78%, and the average classification discrimination than other number of training and test sample number is high.
Embodiment 2
By for extracting 6 kinds of gesture data collection of surface electromyogram signal in UCI databases, acquire 5 ages 20~
The gesture surface electromyogram signal of 22 years old health volunteers, wherein male 2, women 3.It acquires subject and captures article completion
Holding cylinder (Fig. 2 a), portable weight (Fig. 2 b), hand-held flat articles (Fig. 2 c), palm when held object (Fig. 2 d), hold
Sphere (Fig. 2 e), hand pinch the surface electromyogram signal of six kinds of gestures of small article (Fig. 2 f).Each gesture repeats 30 times, everyone adopts
The hand signal of collection is stored with a MAT file, and there are five MAT files altogether, share 900 gesture surface electromyogram signals, including
150 holding cylinders, 150 hand-held flat articles, 150 palm when held objects, are held the puck for 150 times at 150 portable weights
Body, 150 hands pinch small article hand signal.Steps are as follows for surface electromyogram signal gesture identification method:
(1) multistage intrinsic mode function is determined
Surface electromyogram signal gesture data collection is extracted from UCI databases, and table of arbitrary extraction is concentrated from gesture data
Facial muscle electric signal carries out variation mode decomposition, obtains the intrinsic mode function of formula (1):
α is secondary penalty factor in formula,It is the Fourier transformation of gesture surface electromyogram signal,It is that glug is bright
The Fourier transformation of day multiplier,It is the Fourier transformation of i-th of intrinsic mode function,It is k-th natural mode
The Fourier transformation of state function, K 1, n are the newer numbers of expression formula (1), are the number that 0, ω is gesture surface electromyogram signal
Domain frequency, ωk={ ω1,ω2, ,ωkBe whole intrinsic mode functions centre frequency, k is recycled to K from 1, what k was to determine
Mode decomposition number is 1;To intrinsic mode function progress and mode decomposition number phase homogeneous iteration, multistage natural mode of vibration letter is obtained
Number.
It is above-mentioned by gesture data concentrate arbitrary extraction one-time surface electromyography signal carry out variation mode decomposition method and step with
Embodiment 1 is identical.
Other steps are same as Example 1.Obtain gesture identification result.
Embodiment 3
By for extracting 6 kinds of gesture data collection of surface electromyogram signal in UCI databases, acquire 5 ages 20~
The gesture surface electromyogram signal of 22 years old health volunteers, wherein male 2, women 3.It acquires subject and captures article completion
Holding cylinder (Fig. 2 a), portable weight (Fig. 2 b), hand-held flat articles (Fig. 2 c), palm when held object (Fig. 2 d), hold
Sphere (Fig. 2 e), hand pinch the surface electromyogram signal of six kinds of gestures of small article (Fig. 2 f).Each gesture repeats 30 times, everyone adopts
The hand signal of collection is stored with a MAT file, and there are five MAT files altogether, share 900 gesture surface electromyogram signals, including
150 holding cylinders, 150 hand-held flat articles, 150 palm when held objects, are held the puck for 150 times at 150 portable weights
Body, 150 hands pinch small article hand signal.Steps are as follows for surface electromyogram signal gesture identification method:
(1) multistage intrinsic mode function is determined
Surface electromyogram signal gesture data collection is extracted from UCI databases, and table of arbitrary extraction is concentrated from gesture data
Facial muscle electric signal carries out variation mode decomposition, obtains the intrinsic mode function of formula (1):
α is secondary penalty factor in formula,It is the Fourier transformation of gesture surface electromyogram signal,It is that glug is bright
The Fourier transformation of day multiplier,It is the Fourier transformation of i-th of intrinsic mode function,It is k-th natural mode
The Fourier transformation of state function, K 3, n are the newer numbers of expression formula (1), are the number that 2, ω is gesture surface electromyogram signal
Domain frequency, ωk={ ω1,ω2, ,ωkBe whole intrinsic mode functions centre frequency, k is recycled to K from 1, what k was to determine
Mode decomposition number is 1~3;To intrinsic mode function progress and mode decomposition number phase homogeneous iteration, multistage natural mode is obtained
State function.
It is above-mentioned by gesture data concentrate arbitrary extraction one-time surface electromyography signal carry out variation mode decomposition method and step with
Embodiment 1 is identical.
Other steps are same as Example 1.Obtain gesture identification result.
In order to determine that the best-of-breed technology scheme of the present invention, inventor have carried out computer simulation experiment, test situation is as follows:
1, influence of the centre frequency to mode function number K
Table 2, which lists, holds periphery electromyography signal mode function number after the decomposition of variation mode decomposition method
When K takes 1,2,3,4,5, corresponding obtained centre frequency is shown in Table 2.As can be seen from Table 2, when mode function number K is 5, mode letter
Number u2、u3、u4Centre frequency difference be less than 0.1Hz, the centre frequency of three mode functions is close, illustrates mode function number K
When taking 5, decomposition had occurred in signal.As mode function number K<When 4, signal can be caused to owe to decompose, signal occurs to owe to decompose meeting
There are two kinds of situations:The first signal can be added on adjacent mode function after being split and generate aliasing, second of information quilt
It loses;When mode function number K is 4, most preferably, it can decomposite and compare accurately mode function.The present invention chooses mode function
Number K is 1~5, most preferably 4.
The corresponding centre frequencies of 2 different modalities function number K of table
2, influences of the different modalities function number K to gesture Classification and Identification rate
When table 3 is that mode function number K takes 1,2,3,4,5, corresponding gesture classification discrimination is obtained, is shown in Table 3.By table 3
As it can be seen that when K is 4, six kinds of gestures, which are averagely classified, is identified as 93.78%, and gesture when other values is taken averagely to classify discrimination than K
It is high.It is 1~5, most preferably 4 that the present invention, which chooses mode function number K,.
The gesture classification discrimination of 3 different modalities function number K of table
3, influence of the distinct methods decomposed signal to gesture identification rate
Table 4 lists the gesture obtained using overall experience mode decomposition method, variation mode decomposition method decomposed signal
Classification and Identification rate.Surface electromyogram signal is decomposed using variation mode decomposition method, the effect of gesture identification is best, six kinds
The average classification discrimination of gesture is 93.78%.
4 signal of table is by different decomposition method treated discrimination
4, influence of the dimension reduction method to gesture identification rate
Table 5 lists the gesture classification discrimination obtained after different dimension reduction method extraction gesture features.By table 5 as it can be seen that prison
Superintend and direct be locally linear embedding into (SLLE) method extraction feature obtain gesture averagely classify discrimination be 93.78%.SLLE method hands
The average classification discrimination of gesture is higher than the average classification discrimination of other method gestures, and SLLE methods can be with compared with other methods
Obtain more effective gesture feature.The present invention chooses SLLE methods.
The gesture classification discrimination of the different dimension reduction method extraction features of table 5
Claims (3)
1. a kind of surface electromyogram signal gesture identification method, it is characterised in that be made of following step:
(1) multistage intrinsic mode function is determined
Surface electromyogram signal gesture data collection is extracted from UCI databases, and arbitrary extraction one-time surface flesh is concentrated from gesture data
Electric signal carries out variation mode decomposition, obtains the intrinsic mode function of formula (1):
α is penalty factor in formula,It is the Fourier transformation of gesture surface electromyogram signal,It is Lagrange multiplier
Fourier transformation,It is the Fourier transformation of i-th of intrinsic mode function,It is k-th intrinsic mode function
Fourier transformation, n are the newer numbers of expression formula (1), are the numeric field frequency that 0~4, ω is gesture surface electromyogram signal, ωk
={ ω1,ω2,,ωkBe whole intrinsic mode functions centre frequency, k is recycled to K, the mode decomposition number that k is to determine from 1
It is 1~5;To intrinsic mode function progress and mode decomposition number phase homogeneous iteration, multistage intrinsic mode function is obtained;
(2) denoising is carried out to multistage intrinsic mode function
Denoising is carried out using existing Fast Independent Component Analysis method using multistage intrinsic mode function as hand signal;
(3) feature vector of hand signal is obtained
Method is locally linear embedding into using supervision, dimensionality reduction is carried out according to a conventional method to the hand signal after denoising, extracted feature, obtain
To feature vector shown in formula (2):
It is limited positive integer that l, which is the sample number of each gesture, in formula, and i, j are limited positive integer, yiIt is embedded in vector for low-dimensional,
yjIt is yiNeighborhood point, wijIt is weight coefficient;
(4) gesture identification
The feature vector input support vector machine classifier of all surface electromyography signal is concentrated to carry out gesture identification gesture data,
Obtain gesture identification result.
2. surface electromyogram signal gesture identification method according to claim 1, it is characterised in that determining multistage natural mode
It is described to concentrate arbitrary extraction one-time surface electromyography signal to carry out variation mode decomposition from gesture data in state function step (1)
Steps are as follows:
(1) to intrinsic mode function uk(t) Hilbert transform is carried out, unilateral frequency spectrum is obtained:
δ (t) is unit impulse function in formula, and * is convolution, and j is imaginary unit, and t is time, uk(t) it is intrinsic mode function, k is
The number of intrinsic mode function;
(2) the centre frequency ω that index adjusts each modal components is added in each intrinsic mode functionk:
* is convolution in formula, and j is imaginary unit, and t is the time, and δ (t) is unit impulse function.
3. surface electromyogram signal gesture identification method according to claim 1, it is characterised in that:Determining multistage natural mode
In state function step (1), the mode number K is 4, and in the feature vector step (3) for obtaining hand signal, described is every
The sample l of kind gesture is 150.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810552616.8A CN108805067A (en) | 2018-05-31 | 2018-05-31 | Surface electromyogram signal gesture identification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810552616.8A CN108805067A (en) | 2018-05-31 | 2018-05-31 | Surface electromyogram signal gesture identification method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108805067A true CN108805067A (en) | 2018-11-13 |
Family
ID=64089818
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810552616.8A Pending CN108805067A (en) | 2018-05-31 | 2018-05-31 | Surface electromyogram signal gesture identification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108805067A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109934139A (en) * | 2019-03-01 | 2019-06-25 | 浙江工业大学 | A kind of muscle electrical signal paths combined optimization method based on swarm intelligence algorithm |
CN111046731A (en) * | 2019-11-11 | 2020-04-21 | 中国科学院计算技术研究所 | Transfer learning method and recognition method for gesture recognition based on surface electromyogram signals |
CN115034273A (en) * | 2021-12-27 | 2022-09-09 | 驻马店市中心医院 | Myoelectricity biofeedback equipment and system based on pattern recognition |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105446484A (en) * | 2015-11-19 | 2016-03-30 | 浙江大学 | Electromyographic signal gesture recognition method based on hidden markov model |
-
2018
- 2018-05-31 CN CN201810552616.8A patent/CN108805067A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105446484A (en) * | 2015-11-19 | 2016-03-30 | 浙江大学 | Electromyographic signal gesture recognition method based on hidden markov model |
Non-Patent Citations (4)
Title |
---|
刘长良 等: "基于变分模态分解和模糊C均值聚类的滚动轴承故障诊断", 《中国电机工程学报》 * |
杜义浩 等: "基于变分模态分解_相干分析的肌间耦合特性", 《物理学报》 * |
栗科峰: "《盲信号处理技术及工程应用》", 30 November 2017, 黄河水利出版社 * |
马欣欣 等: "基于EEMD和多域特征融合的手势肌电信号识别研究", 《云南大学学报(自然科学版)》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109934139A (en) * | 2019-03-01 | 2019-06-25 | 浙江工业大学 | A kind of muscle electrical signal paths combined optimization method based on swarm intelligence algorithm |
CN111046731A (en) * | 2019-11-11 | 2020-04-21 | 中国科学院计算技术研究所 | Transfer learning method and recognition method for gesture recognition based on surface electromyogram signals |
CN111046731B (en) * | 2019-11-11 | 2023-07-25 | 中国科学院计算技术研究所 | Transfer learning method and recognition method for gesture recognition based on surface electromyographic signals |
CN115034273A (en) * | 2021-12-27 | 2022-09-09 | 驻马店市中心医院 | Myoelectricity biofeedback equipment and system based on pattern recognition |
CN115034273B (en) * | 2021-12-27 | 2023-09-01 | 驻马店市中心医院 | Myoelectricity biofeedback equipment and system based on pattern recognition |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Lu et al. | A study of personal recognition method based on EMG signal | |
CN105446484B (en) | A kind of electromyography signal gesture identification method based on Hidden Markov Model | |
CN109924977A (en) | A kind of surface electromyogram signal classification method based on CNN and LSTM | |
CN111407243B (en) | Pulse signal pressure identification method based on deep learning | |
CN103440498A (en) | Surface electromyogram signal identification method based on LDA algorithm | |
CN102622605A (en) | Surface electromyogram signal feature extraction and action pattern recognition method | |
CN108805067A (en) | Surface electromyogram signal gesture identification method | |
Zhao et al. | Deep CNN model based on serial-parallel structure optimization for four-class motor imagery EEG classification | |
CN111563581B (en) | Brain muscle function network construction method based on wavelet coherence | |
Sepahvand et al. | A novel multi-lead ECG personal recognition based on signals functional and structural dependencies using time-frequency representation and evolutionary morphological CNN | |
Oweis et al. | ANN-based EMG classification for myoelectric control | |
CN109598222A (en) | Wavelet neural network Mental imagery brain electricity classification method based on the enhancing of EEMD data | |
Naik et al. | Classification of finger extension and flexion of EMG and Cyberglove data with modified ICA weight matrix | |
CN107122643A (en) | Personal identification method based on PPG signals and breath signal Fusion Features | |
CN107088069A (en) | Personal identification method based on human body PPG signal subsections | |
CN110811633A (en) | Identity recognition method, system and device based on electromyographic signals | |
Mahdavi et al. | Surface electromyography feature extraction based on wavelet transform | |
CN108776783A (en) | The gesture electromyography signal recognition methods of extreme learning machine-Hidden Markov Model | |
Oleinikov et al. | Feature extraction and real-time recognition of hand motion intentions from EMGs via artificial neural networks | |
Narayan | Direct comparison of SVM and LR classifier for SEMG signal classification using TFD features | |
Bai et al. | Multi-channel sEMG signal gesture recognition based on improved CNN-LSTM hybrid models | |
CN113749656A (en) | Emotion identification method and device based on multi-dimensional physiological signals | |
Guo et al. | A wavelet packet based pulse waveform analysis for cholecystitis and nephrotic syndrome diagnosis | |
Vigneshwari et al. | Analysis of finger movements using EEG signal | |
Yu et al. | The research of sEMG movement pattern classification based on multiple fused wavelet function |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181113 |