CN108985157A - A kind of gesture identification method and device - Google Patents
A kind of gesture identification method and device Download PDFInfo
- Publication number
- CN108985157A CN108985157A CN201810582318.3A CN201810582318A CN108985157A CN 108985157 A CN108985157 A CN 108985157A CN 201810582318 A CN201810582318 A CN 201810582318A CN 108985157 A CN108985157 A CN 108985157A
- Authority
- CN
- China
- Prior art keywords
- electromyography signal
- gesture
- identification method
- model
- gesture identification
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
Abstract
The invention discloses a kind of gesture identification methods, and the method includes the following steps: by multiple channels acquire target gesture object electromyography signal, the electromyography signal be contraction of muscle when muscle fibre in moving cell one-dimensional time series;Effective electromyography signal is extracted according to the period of motion of default gesture;By in effective electromyography signal input linear prediction autoregression AR model, characteristic vector pickup is carried out, the described eigenvector that multiple channels are extracted is merged, multidimensional characteristic vectors are combined into;Described eigenvector is inputted in the K-means model of preset standardized centroid cluster, the cluster similarity of each mass center, feature vector is divided into the mass center of highest similarity, to identify to gesture motion in analysis feature vector and the K-means model.The invention also discloses a kind of devices based on above-mentioned gesture identification method.
Description
Technical field
The present invention relates to technical field of hand gesture recognition, a kind of gesture identification method and device are particularly related to.
Background technique
In recent years, computer application technology with various human-computer interaction sensors emergence and development, in addition to tradition
Keyboard and mouse input other than, emerged more kinds of man-machine interaction modes.Wherein based on the human-computer interaction of gesture identification
Mode is widely deployed because of its intuitive, simplicity, the features such as operation is convenient in fields such as rehabilitation, Mechanical courses.At present
Gesture identification method be broadly divided into identification method based on computer vision, the identification method based on pressure signal, based on spiral shell
Identification method, identification method based on EMG (electromyography signal) for revolving instrument sensor etc..
Common identification method on the market, as Microsoft Research propose Handpose (Gesture Recognition Algorithm) software,
Images of gestures is identified using Kinect (dynamics connection) camera, restores its gesture motion, but such method without
Method solves the problems such as gesture masking, camera blind area, and can not move, inconvenient for use.And known using gyroscope sensor
Other method can only classify to the movement that amplitude changes greatly, can not identify to fine movement.
As it can be seen that there are many problems for current gesture identification method.Therefore, a kind of new gesture identification method is needed, is come
Solve that existing gesture identification method is cumbersome, the not high problem of the accuracy of identification to gesture.
Summary of the invention
In view of this, it is an object of the invention to propose the accuracy of identification of a kind of pair of gesture high gesture identification method and dress
It sets.
Based on a kind of above-mentioned purpose gesture identification method provided by the invention, the method includes the following steps:
The electromyography signal of target gesture object is acquired by multiple channels, when the electromyography signal is contraction of muscle in muscle fibre
The one-dimensional time series of moving cell;
Effective electromyography signal is extracted according to the period of motion of default gesture;
By in effective electromyography signal input linear prediction autoregression AR model, characteristic vector pickup is carried out, it will be multiple
The described eigenvector that the channel is extracted is merged, and multidimensional characteristic vectors are combined into;
Described eigenvector is inputted in the K-means model of preset standardized centroid cluster, analysis feature vector with it is described
The cluster similarity of each mass center, feature vector is divided into the mass center of highest similarity in K-means model, with dynamic to gesture
It is identified.
It includes: the linear pre- of acquisition effective electromyography signal that described eigenvector, which is extracted, in one of the embodiments,
Coefficient value is surveyed, and according to the linear predictor coefficient value, extracts one group of data that can characterize the electromyography signal inherent characteristic.
The linear prediction AR model is in one of the embodiments,
Wherein, x (n) is the current time value of electromyography signal, and w (n) is white noise, and p is prediction model order, akFor feature vector, x
It (n-k) is the value at electromyography signal k moment more early than current time.
The K-means model is in one of the embodiments,Wherein, V is to miss
Poor quadratic sum, k are k cluster centre, μiFor cluster centre point, xjFor target point, SiFor gathering.
The building of the K-means model of the preset standardized centroid cluster includes: that will preset in one of the embodiments,
The standard electromyography signal of standard gesture motion be divided into several clusters, each standard electromyography signal is similar in the cluster
Degree is calculated according to the mean value mass center obtained of all standard electromyography signals.
The period of motion of the default gesture is 2000ms in one of the embodiments,.
The multiple channel is 8 in one of the embodiments, and each channel includes an electrode.
In one of the embodiments, after the electromyography signal of the acquisition target gesture object, effective electromyography signal is extracted
Before further include: the electromyography signal is pre-processed.
The pretreatment in one of the embodiments, includes, by bandpass filtering or bandreject filtering, removing myoelectricity letter
Ambient noise in number.
The present invention also provides a kind of devices applied to above-mentioned gesture identification method, comprising:
Electromyographic signal collection module, for acquiring the electromyography signal of target gesture object, the myoelectricity letter by multiple channels
Number be contraction of muscle when muscle fibre in moving cell one-dimensional time series;
Effective electromyography signal extraction module, for extracting effective electromyography signal according to the period of motion of default gesture;
Characteristic vector pickup module, for effective electromyography signal input linear to be predicted in autoregression AR model, into
The characteristic that multiple channels are extracted is merged, is combined into multidimensional characteristic vectors by row characteristic vector pickup;
Eigenvector recognition module, for described eigenvector to be inputted to the K-means model of preset standardized centroid cluster
In, the cluster similarity of each mass center, it is similar to be divided into highest for feature vector in analysis feature vector and the K-means model
In the mass center of degree, to be identified to gesture motion.
From the above it can be seen that the gesture identification method provided by the invention based on electromyography signal, passes through multichannel
It acquires and sends the electromyography signal of target gesture object to local server, for effective electromyography signal of extraction, by linear pre-
The characteristics of survey autoregression AR model extraction feature vector, the models coupling surface electromyogram signal, by the random of surface electromyogram signal
Property and predictability combine, and stable electromyography signal feature vector in a short time can be extracted, to improve feature
The accuracy that vector extracts;And the K-means model of preset standardized centroid cluster, the mould are inputted after combining described eigenvector
Type is filtered out highest with feature vector similarity by the similarity of each mass center in analysis feature vector and K-means model
Mass center, and feature vector is divided to wherein, to carry out accurate analysis and identification to feature vector, realize to gesture motion
What is carried out accurately identifies.
Detailed description of the invention
Fig. 1 is the flow chart of the gesture identification method based on electromyography signal of the embodiment of the present invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with specific embodiment, and reference
Attached drawing, the present invention is described in more detail.
Referring to Fig. 1, the present invention provides a kind of gesture identification method based on electromyography signal, the method includes following steps
It is rapid:
S100 acquires the electromyography signal of target gesture object, the flesh when electromyography signal is contraction of muscle by multiple channels
The one-dimensional time series of moving cell in fiber.
Specifically, in step S100, the multiple channel can be 8, and each channel all has an electrode.
Preferably, it after step S100, may also include, the electromyography signal pre-processed.It, can be with by pretreatment
The interference signal in electromyography signal is effectively rejected, the efficiency to the subsequent processing of electromyography signal is improved.
The pretreatment may include, by bandpass filtering or bandreject filtering, removing the ambient noise in electromyography signal.
S200 extracts effective electromyography signal according to the period of motion of default gesture.
The period of motion of the default gesture is 2000ms.
S300 carries out characteristic vector pickup in effective electromyography signal input linear prediction autoregression AR model, will
The described eigenvector that multiple channels are extracted is merged, and multidimensional characteristic vectors are combined into.
Described eigenvector extraction includes: to obtain the linear predictor coefficient value of effective electromyography signal, and according to described
Linear predictor coefficient value extracts one group of data that can characterize the electromyography signal inherent characteristic.
The linear prediction autoregression AR model isWherein, x (n) is myoelectricity
The current time value of signal, w (n) are white noise, and p is prediction model order, akFor feature vector, x (n-k) is electromyography signal ratio
The value at k moment of morning at current time.
The AR model, can effectively mating surface electromyography signal the characteristics of, by the randomness of surface electromyogram signal and can be pre-
The property surveyed combines.EMG (electromyography signal) can be considered as the generation of some determination system of white-noise excitation by the model, and be led to
The output input relationship of research exciting power value and model parameter value research system is crossed, and then studies EMG characteristic, therefore can mention
Take stable electromyography signal feature vector in a short time.The structure feature of the AR model carries out parameter using least square method and estimates
Meter, to improve computational accuracy while reducing calculation amount to the greatest extent, has the advantages that simple and easy.
Specifically, p can be 5, be the current value of list entries and the related coefficient of first five moment value.At this point, each
There are 5 feature vectors in a channel.
S400 inputs described eigenvector in the K-means model of preset standardized centroid cluster, analysis feature vector with
The cluster similarity of each mass center, feature vector is divided into the mass center of highest similarity, with opponent in the K-means model
Gesture movement is identified.
The building of the K-means model of the preset standardized centroid cluster includes: by the mark of preset standard gesture motion
Quasi- electromyography signal is divided into several clusters, and the similarity of each standard electromyography signal is according to all standard myoelectricities in the cluster
The mean value of signal mass center obtained is calculated.
The K-means model isWherein, V is error sum of squares, and k is k cluster
Center, μiFor cluster centre point, xjFor target point, S is gathering.
In the K-means model of the preset standardized centroid cluster, in several clusters, each standard flesh in same cluster
The similarity of electric signal is higher, and the similarity of each standard electromyography signal is lower in different clusters.
The gesture motion of Classification and Identification of the present invention is mainly to cover the activity of wrist, arm part, including stretching, extension of clenching fist, arm
It swings, twisting, wrist or so twisting, four refer to that bending, thumb are bent seven and act up and down for arms swing up downward, wrist.
Gesture identification method provided by the invention based on electromyography signal passes through multichannel collecting and sends target gesture object
Electromyography signal pass through linear prediction autoregression AR model extraction feature vector, the model for effective electromyography signal of extraction
The characteristics of mating surface electromyography signal, combines the randomness of surface electromyogram signal and predictability, can extract short
Stable electromyography signal feature vector in time, to improve the accuracy of characteristic vector pickup;And by described eigenvector
The K-means model of preset standardized centroid cluster is inputted after combination, which passes through in analysis feature vector and K-means model
The similarity of each mass center, filter out with the highest mass center of feature vector similarity, and feature vector is divided to wherein, thus right
Feature vector carries out accurate analysis and identification, realizes and identifies to the precise classification that gesture motion carries out.
Gesture identification method based on electromyography signal of the invention can be applied to the movement of the rehabilitation training of disability crowd
It counts.
The present invention also provides a kind of devices applied to the above-mentioned gesture identification method based on electromyography signal, comprising:
Electromyographic signal collection module, for acquiring the electromyography signal of target gesture object, the myoelectricity letter by multiple channels
Number be contraction of muscle when muscle fibre in moving cell one-dimensional time series.
The electromyographic signal collection module includes sensing unit, and the sensing unit has multiple channels, each described logical
Road includes an electrode.After sensing unit acquires electromyography signal, local server can be sent to by equipment such as bluetooths.This
Ground server can be the equipment such as computer, tablet computer or mobile phone.
The sensing unit is the armlet that the fore-arm of object to be measured is arranged in.Armlet is specifically as follows preferably, armlet
It is arranged at the brachioradialis of the forethiga of the object to be measured, it, will not be to object to be measured on activity at wrist, finger without influence
Gesture mobility cause to limit, so that the flexibility of gesture operation be greatly improved.
It can also include preprocessing module, for pre-processing to electromyography signal, the environment removed in electromyography signal is made an uproar
Sound.
Effective electromyography signal extraction module extracts the movement for meeting default gesture for the period of motion according to default gesture
Effective electromyography signal in period.
Characteristic vector pickup module, for effective electromyography signal input linear to be predicted in autoregression AR model, into
The characteristic that multiple channels are extracted is merged, is combined into multidimensional characteristic vectors by row characteristic vector pickup.
Eigenvector recognition module, for described eigenvector to be inputted to the K-means model of preset standardized centroid cluster
In, the cluster similarity of each mass center, it is similar to be divided into highest for feature vector in analysis feature vector and the K-means model
In the mass center of degree, thus to be identified to gesture motion.
In actual use, user can wear MYO wrist strap (gesture control armlet) at forethiga brachioradialis, carry out
After movement, MYO armlet acquires the EMG (electromyography signal) of user, is sent to computer by bluetooth, computer extracts user
After effective electromyography signal of one action cycle, inputs in AR model, obtain the feature vector of 40 dimensions, then this feature vector is defeated
Enter in K-means model, is compared with each standardized centroid cluster in model, calculate its similarity (Euclidean distance) with every cluster, it is defeated
The mass center cluster serial number of similarity maximum (Euclidean distance is minimum) is as differentiation as a result, being used as corresponding gesture motion out.It please join
Table 1 is read, it is right when being applied to the rehabilitation training of disability crowd for the gesture identification method and device of the invention based on electromyography signal
The accuracy rate of the recognition result of the movement of disabled.Wherein, it is that arm swings movement, M3 that M1, which is clench fist stretching, M2,
It is acted for arms swing up downward, M4 is wrist or more twisting actions, M5 is wrist or so twisting actions, M6 is that the bending of four fingers is dynamic
Make, M7 is thumb flexure operation.
1 gesture identification accuracy rate of table
It can be seen that the gesture identification method and device of the invention based on electromyography signal, can be very good dynamic to gesture
Make carry out Classification and Identification, and accuracy rate is high, coverage area is big, can cover human body large arm, forearm, and the gesture activity of the five fingers has
The efficiency of excellent gesture identification.
It should be understood by those ordinary skilled in the art that: the discussion of any of the above embodiment is exemplary only, not
It is intended to imply that the scope of the present disclosure (including claim) is limited to these examples;Under thinking of the invention, above embodiments
Or can also be combined between the technical characteristic in different embodiments, step can be realized with random order, and be existed such as
Many other variations of the upper different aspect of the invention, for simplicity, they are not provided in details.
Although having been incorporated with specific embodiments of the present invention, invention has been described, according to retouching for front
It states, many replacements of these embodiments, modifications and variations will be apparent for those of ordinary skills.Example
Such as, discussed embodiment can be used in other memory architectures (for example, dynamic ram (DRAM)).
The embodiment of the present invention be intended to cover fall into all such replacements within the broad range of appended claims,
Modifications and variations.Therefore, all within the spirits and principles of the present invention, any omission, modification, equivalent replacement, the improvement made
Deng should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of gesture identification method, which is characterized in that the method includes the following steps:
The electromyography signal of target gesture object is acquired by multiple channels, is moved in muscle fibre when the electromyography signal is contraction of muscle
The one-dimensional time series of unit;
Effective electromyography signal is extracted according to the period of motion of default gesture;
By in effective electromyography signal input linear prediction autoregression AR model, characteristic vector pickup is carried out, it will be multiple described
The described eigenvector that channel is extracted is merged, and multidimensional characteristic vectors are combined into;
Described eigenvector is inputted in the K-means model of preset standardized centroid cluster, analysis feature vector and the K-
The cluster similarity of each mass center, feature vector is divided into the mass center of highest similarity, to gesture motion in means model
It is identified.
2. gesture identification method according to claim 1, which is characterized in that described eigenvector extraction includes: to obtain institute
The linear predictor coefficient value of effective electromyography signal is stated, and according to the linear predictor coefficient value, the flesh can be characterized by extracting one group
The data of electric signal inherent characteristic.
3. gesture identification method according to claim 1, which is characterized in that the linear prediction AR model isWherein, x (n) is the current time value of electromyography signal, and w (n) is white noise, and p is
Prediction model order, akFor feature vector, x (n-k) is the value at electromyography signal k moment more early than current time.
4. gesture identification method according to claim 1, which is characterized in that the K-means model isWherein, V is error sum of squares, and k is k cluster centre, μiFor cluster centre point, xjFor mesh
Punctuate, SiFor gathering.
5. gesture identification method according to claim 1, which is characterized in that the K- of the preset standardized centroid cluster
The building of means model includes: that the standard electromyography signal of preset standard gesture motion is divided into several clusters, described poly-
The similarity of each standard electromyography signal is calculated according to the mean value mass center obtained of all standard electromyography signals in class.
6. gesture identification method according to claim 1, which is characterized in that the period of motion of the default gesture is
2000ms。
7. gesture identification method according to claim 1, which is characterized in that the multiple channel is 8, each described logical
Road includes an electrode.
8. gesture identification method according to claim 1, which is characterized in that the electromyography signal of the acquisition target gesture object
Later, before extracting effective electromyography signal further include: pre-processed to the electromyography signal.
9. gesture identification method according to claim 8, which is characterized in that the pretreatment includes passing through bandpass filtering
Or bandreject filtering, remove the ambient noise in electromyography signal.
10. a kind of device applied to the described in any item gesture identification methods of claim 1 to 9 characterized by comprising
Electromyographic signal collection module, for acquiring the electromyography signal of target gesture object by multiple channels, the electromyography signal is
When contraction of muscle in muscle fibre moving cell one-dimensional time series;
Effective electromyography signal extraction module, for extracting effective electromyography signal according to the period of motion of default gesture;
Characteristic vector pickup module, for carrying out in effective electromyography signal input linear prediction autoregression AR model special
It levies vector to extract, the characteristic that multiple channels are extracted is merged, multidimensional characteristic vectors are combined into;
Eigenvector recognition module, for described eigenvector being inputted in the K-means model of preset standardized centroid cluster, point
The cluster similarity for analysing each mass center in feature vector and the K-means model, is divided into highest similarity for feature vector
In mass center, to be identified to gesture motion.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810582318.3A CN108985157A (en) | 2018-06-07 | 2018-06-07 | A kind of gesture identification method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810582318.3A CN108985157A (en) | 2018-06-07 | 2018-06-07 | A kind of gesture identification method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108985157A true CN108985157A (en) | 2018-12-11 |
Family
ID=64540993
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810582318.3A Pending CN108985157A (en) | 2018-06-07 | 2018-06-07 | A kind of gesture identification method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108985157A (en) |
Cited By (7)
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 |
CN110163086A (en) * | 2019-04-09 | 2019-08-23 | 缤刻普达(北京)科技有限责任公司 | Body-building action identification method, device, equipment and medium neural network based |
CN110568929A (en) * | 2019-09-06 | 2019-12-13 | 诺百爱(杭州)科技有限责任公司 | Virtual scene interaction method and device and electronic equipment |
CN110618754A (en) * | 2019-08-30 | 2019-12-27 | 电子科技大学 | Surface electromyogram signal-based gesture recognition method and gesture recognition armband |
CN111371951A (en) * | 2020-03-03 | 2020-07-03 | 北京航空航天大学 | Smart phone user authentication method and system based on electromyographic signals and twin neural network |
CN112328076A (en) * | 2020-11-06 | 2021-02-05 | 北京中科深智科技有限公司 | Method and system for driving character gestures through voice |
CN112597820A (en) * | 2020-12-10 | 2021-04-02 | 南京长峰航天电子科技有限公司 | Target clustering method based on radar signal sorting |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105326501A (en) * | 2015-12-10 | 2016-02-17 | 宁波工程学院 | Muscle disease monitoring method based on sEMG |
CN105426842A (en) * | 2015-11-19 | 2016-03-23 | 浙江大学 | Support vector machine based surface electromyogram signal multi-hand action identification method |
CN105446484A (en) * | 2015-11-19 | 2016-03-30 | 浙江大学 | Electromyographic signal gesture recognition method based on hidden markov model |
US20160180644A1 (en) * | 2014-12-17 | 2016-06-23 | Fayez Idris | Gaming system with movable ultrasonic transducer |
CN106845348A (en) * | 2016-12-20 | 2017-06-13 | 南京信息工程大学 | A kind of gesture identification method based on arm surface electromyographic signal |
-
2018
- 2018-06-07 CN CN201810582318.3A patent/CN108985157A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160180644A1 (en) * | 2014-12-17 | 2016-06-23 | Fayez Idris | Gaming system with movable ultrasonic transducer |
CN105426842A (en) * | 2015-11-19 | 2016-03-23 | 浙江大学 | Support vector machine based surface electromyogram signal multi-hand action identification method |
CN105446484A (en) * | 2015-11-19 | 2016-03-30 | 浙江大学 | Electromyographic signal gesture recognition method based on hidden markov model |
CN105326501A (en) * | 2015-12-10 | 2016-02-17 | 宁波工程学院 | Muscle disease monitoring method based on sEMG |
CN106845348A (en) * | 2016-12-20 | 2017-06-13 | 南京信息工程大学 | A kind of gesture identification method based on arm surface electromyographic signal |
Non-Patent Citations (1)
Title |
---|
刘建伟: "肌电控制接口的自适应方法研究", 《万方学术论文在线平台》 * |
Cited By (11)
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 |
CN110163086A (en) * | 2019-04-09 | 2019-08-23 | 缤刻普达(北京)科技有限责任公司 | Body-building action identification method, device, equipment and medium neural network based |
CN110618754A (en) * | 2019-08-30 | 2019-12-27 | 电子科技大学 | Surface electromyogram signal-based gesture recognition method and gesture recognition armband |
CN110618754B (en) * | 2019-08-30 | 2021-09-14 | 电子科技大学 | Surface electromyogram signal-based gesture recognition method and gesture recognition armband |
CN110568929A (en) * | 2019-09-06 | 2019-12-13 | 诺百爱(杭州)科技有限责任公司 | Virtual scene interaction method and device and electronic equipment |
CN110568929B (en) * | 2019-09-06 | 2023-04-25 | 诺百爱(杭州)科技有限责任公司 | Virtual scene interaction method and device based on virtual keyboard and electronic equipment |
CN111371951A (en) * | 2020-03-03 | 2020-07-03 | 北京航空航天大学 | Smart phone user authentication method and system based on electromyographic signals and twin neural network |
CN111371951B (en) * | 2020-03-03 | 2021-04-23 | 北京航空航天大学 | Smart phone user authentication method and system based on electromyographic signals and twin neural network |
CN112328076A (en) * | 2020-11-06 | 2021-02-05 | 北京中科深智科技有限公司 | Method and system for driving character gestures through voice |
CN112328076B (en) * | 2020-11-06 | 2021-10-29 | 北京中科深智科技有限公司 | Method and system for driving character gestures through voice |
CN112597820A (en) * | 2020-12-10 | 2021-04-02 | 南京长峰航天电子科技有限公司 | Target clustering method based on radar signal sorting |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108985157A (en) | A kind of gesture identification method and device | |
Yang et al. | Chinese sign language recognition based on an optimized tree-structure framework | |
Kerber et al. | User-independent real-time hand gesture recognition based on surface electromyography | |
CN106919958B (en) | Human body finger action recognition method based on smart watch | |
CN110555468A (en) | Electroencephalogram signal identification method and system combining recursion graph and CNN | |
TW202036245A (en) | System for recognizing user gestures according to mechanomyogram detected from user’s wrist and method thereof | |
CN101776981B (en) | Method for controlling mouse by jointing brain electricity and myoelectricity | |
CN107145236B (en) | Gesture recognition method and system based on wrist tendon pressure related characteristics | |
CN110826625B (en) | Finger gesture classification method based on surface electromyographic signals | |
Zhang et al. | Recognizing hand gestures with pressure-sensor-based motion sensing | |
CN108196668B (en) | Portable gesture recognition system and method | |
Bian et al. | SVM based simultaneous hand movements classification using sEMG signals | |
Shin et al. | Korean sign language recognition using EMG and IMU sensors based on group-dependent NN models | |
CN112464738B (en) | Improved naive Bayes algorithm user behavior identification method based on mobile phone sensor | |
CN107480697A (en) | A kind of myoelectricity gesture identification method and system | |
Fahim et al. | A visual analytic in deep learning approach to eye movement for human-machine interaction based on inertia measurement | |
CN109765996A (en) | Insensitive gesture detection system and method are deviated to wearing position based on FMG armband | |
Fan et al. | Wearable motion attitude detection and data analysis based on Internet of Things | |
CN111700718A (en) | Holding posture identifying method, holding posture identifying device, artificial limb and readable storage medium | |
CN110532898A (en) | A kind of physical activity recognition methods based on smart phone Multi-sensor Fusion | |
Yu et al. | Real-time multiple gesture recognition: application of a lightweight individualized 1D CNN model to an edge computing system | |
Geng | Research on athlete’s action recognition based on acceleration sensor and deep learning | |
Lv et al. | Common spatial pattern and particle swarm optimization for channel selection in BCI | |
CN110751060B (en) | Portable motion mode real-time identification system based on multi-source signals | |
Chunsheng et al. | Human gait feature data analysis and person identification based on IMU |
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 |
Application publication date: 20181211 |
|
RJ01 | Rejection of invention patent application after publication |