CN110389663A - A kind of sEMG gesture identification method based on small wave width learning system - Google Patents
A kind of sEMG gesture identification method based on small wave width learning system Download PDFInfo
- Publication number
- CN110389663A CN110389663A CN201910548939.4A CN201910548939A CN110389663A CN 110389663 A CN110389663 A CN 110389663A CN 201910548939 A CN201910548939 A CN 201910548939A CN 110389663 A CN110389663 A CN 110389663A
- Authority
- CN
- China
- Prior art keywords
- semg
- signal
- wave width
- small wave
- gesture
- 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.)
- Granted
Links
Classifications
-
- 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
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
- G06F3/015—Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/017—Gesture based interaction, e.g. based on a set of recognized hand gestures
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Computer Interaction (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Multimedia (AREA)
- Biomedical Technology (AREA)
- Dermatology (AREA)
- Neurology (AREA)
- Neurosurgery (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of sEMG gesture identification methods based on small wave width learning system, include the following steps: step 1, and the quantity d of gesture motion type is defined according to corresponding identification scene;Step 2 acquires sEMG signal s using electromyographic signal collection equipment;Step 3 is filtered noise reduction to sEMG original signal, while removing the noise other than 10Hz~500Hz frequency band using butterworth filter according to the frequency characteristic of sEMG signal;Step 4, using the active segment in Moving Window method detection sEMG signal;Step 5 carries out feature extraction to the active segment that detection obtains;The present invention can be more quickly completed the training of model and the determination of parameter compared to the algorithm based on conventional depth learning network, to improve working efficiency;Node is expanded dynamically to improve the discrimination of system, without re-establishing completely and training pattern.
Description
Technical field
The present invention relates to machine learning and Modulation recognition technical field, and in particular to one kind is based on small wave width learning system
SEMG gesture identification method.
Background technique
Gesture motion in human body language plays an important role among daily exchange, such as the movement of football mat chairman
Etc..Therefore, how scholars allow computer and machine that can efficiently, accurately identify what the mankind issued with regard to primary study
Gesture motion simultaneously executes corresponding program.This will change the form of communication of the mankind and machine.
Currently, being broadly divided into for the recognizer of gesture motion following several: the classification of view-based access control model image recognition is calculated
Method and the Gesture Recognition Algorithm for being based on surface electromyogram signal (sEMG);The former is handled for image, and this method is to camera shooting
The requirement of equipment is relatively high while the price of equipment also can be more expensive, and therefore, it is difficult to universal, but discrimination has been studied
Higher discrimination;And it is lower to hardware requirement compared to for the former based on the Gesture Recognition Algorithm of surface electromyogram signal (sEMG)
And it is easy to accomplish, while the hardware for acquiring sEMG signal can wear and be also not limited to fixed camera position
Identify product.Current existing method is much to go how research classifies to signal based on intelligent algorithm.
In recent years, the largely research about gesture motion recognizer has been carried out both at home and abroad, comprising: signal acquisition is ground
Study carefully, to the research of feature extraction and calculating, the research of signal filtering and research of tagsort model etc..Currently, common
Disaggregated model be all based on classical machine learning algorithm, such as random forest network, BP neural network, convolutional neural networks
Etc..But these classical network models can all consume a large amount of time in the training process.
Summary of the invention
The purpose of the present invention is to overcome the shortcomings of the existing technology and deficiency, provides a kind of based on small wave width learning system
SEMG gesture identification method, this method can be more quickly completed the training of model and the determination of parameter, to improve work
Efficiency.
The purpose of the invention is achieved by the following technical solution:
A kind of sEMG gesture identification method based on small wave width learning system, includes the following steps:
Step 1 defines the quantity d of gesture motion type according to corresponding identification scene, and to each gesture distribution one
A serial number;
Step 2 acquires sEMG signal s using electromyographic signal collection equipment;
Step 3 is filtered noise reduction to sEMG original signal, while utilizing Bart according to the frequency characteristic of sEMG signal
Fertile hereby filter removes the noise other than 10Hz~500Hz frequency band:
Wherein N is the order of filter, ωcFor cutoff frequency filter;
Step 4, using the active segment in Moving Window method detection sEMG signal;
Step 5 carries out feature extraction to the active segment that detection obtains;The feature of signal segment is calculated, this feature includes: flat
Equal absolute value q1, root mean square q2, median frequency q3, means frequency q4..., each feature for active segment of connecting is feature vector:
xk=[q1,q2,...,qn]T,
Wherein k is kth section active signal section;
Further, using the feature vector of each active segment as the input vector x of small wave width learning classifier system;
Step 6, the quantitative design of feature small wave width study system in the gesture quantity and step 5 defined according to step 1
The Inport And Outport Node of system, and the feature vector obtained in step 5 is inputted into small wave width learning system and is classified;It is instructing
According to the Sample Refreshment parameter for having label during white silk;In test and practical process, from the small wave width learning system obtained
The serial number of highest incentive degree node is judged in several nodes of output, then the serial number of the node is the serial number of classification movement;
Step 7 can be known according to the corresponding serial number of each gesture is defined in step 1 by the serial number of classification results
Not Chu gesture type.
Preferably, the Moving Window method in the step 4 specifically:
The signal data in a bit of time is extracted first and carries out the operation of integrated square, as shown in following formula:
Wherein s (t) is the electromyography signal data in window, SiIndicate tiThe energy value at the integral of time-ofday signals i.e. corresponding moment;
A threshold value beta is set to energy value SiIt makes a decision, works as tiThe energy value S at momentiGreater than the continuous n after threshold value beta and moving window1
Secondary energy value is all larger than threshold value, then it is assumed that moment tiAt the beginning of movement;On this basis, when from tjThe energy at moment
Value SjLess than threshold value beta and there is continuous n2Secondary energy value is less than threshold value and then thinks tjTo act finish time, therefore, action signal section
Then are as follows:
Wherein k is the movement that kth section detected.
Preferably, small wave width learning system in the step 6 specifically:
(1) different groups of characteristic nodes are mapped using input vector:
Wherein wi, aiAnd biWeight, transfer parameters and the scaling parameter respectively mapped, these parameters are in initialization
It is random to generate and updated using k-mean algorithm;φi() is i-th of wavelet basis function, and wherein n is that wavelet basis function is always a
Number;
(2) series connection each group characteristic node:
Zn=[Z1,Z2,......,Zn];
(3) characteristic node Z is utilizedn, map incremental nodes:
Hm=ξ (ZnWh+βh),
Wherein WhAnd βhThe respectively weight and threshold parameter of incremental nodes, this partial parameters generate at random in initialization
Just fixation does not need to update;ξ () is general excitation function, and sigmoid function may be used herein;
(4) in training process, the weighting parameter of output is obtained using pseudo- reciprocal value and ridge regression algorithm:
Wall=[Zn|Hm]+Y,
Wherein Y is the reference output of training set, and [Zn|Hm]+For [Zn|Hm] pseudo- reciprocal value;In test and practical process,
Output node can directly be mapped out
The present invention have compared with prior art it is below the utility model has the advantages that
The present invention can be more quickly completed the training and parameter of model compared to the algorithm based on conventional depth learning network
It determines, to improve working efficiency;Node is dynamically expanded to improve the discrimination of system, without building again completely
Vertical and training pattern.
Detailed description of the invention
Fig. 1 is flow diagram of the invention;
Fig. 2 is the structural schematic diagram of the small wave width learning system of the present invention;
Fig. 3 is electromyographic signal collection equipment of the present invention (Myo armlet).
Specific embodiment
Present invention will now be described in further detail with reference to the embodiments and the accompanying drawings, but embodiments of the present invention are unlimited
In this.
As shown in Figures 1 to 3, a kind of sEMG gesture identification method based on small wave width learning system, includes the following steps:
Step 1 defines the quantity d of gesture motion type according to corresponding identification scene, and to each gesture distribution one
A serial number.
Step 2 acquires sEMG signal s using electromyographic signal collection equipment as shown in Figure 3.
Step 3 is filtered noise reduction to sEMG original signal, while utilizing Bart according to the frequency characteristic of sEMG signal
Fertile hereby filter removes the noise other than 10Hz~500Hz frequency band:
Wherein N is the order of filter, ωcFor cutoff frequency filter.
Step 4, using the active segment in Moving Window method detection sEMG signal;
The Moving Window method specifically:
The signal data in a bit of time is extracted first and carries out the operation of integrated square, as shown in following formula:
Wherein s (t) is the electromyography signal data in window, SiIndicate tiThe energy value at the integral of time-ofday signals i.e. corresponding moment;
A threshold value beta is set to energy value SiIt makes a decision, works as tiThe energy value S at momentiGreater than the continuous n after threshold value beta and moving window1
Secondary energy value is all larger than threshold value, then it is assumed that moment tiAt the beginning of movement;On this basis, when from tjThe energy at moment
Value SjLess than threshold value beta and there is continuous n2Secondary energy value is less than threshold value and then thinks tjTo act finish time, therefore, action signal section
Then are as follows:
Wherein k is the movement that kth section detected.
Step 5, the active segment signal segment that detection is obtainedCarry out feature extraction;Calculate the feature of signal segment, the spy
Sign includes: average absolute value q1, root mean square q2, median frequency q3, means frequency q4..., each feature for active segment of connecting is
Feature vector:
xk=[q1,q2,...,qn]T,
Wherein k is kth section active signal section;
Further, using the feature vector of each active segment as the input vector x of small wave width learning classifier system.
Step 6, the quantitative design of feature small wave width study system in the gesture quantity and step 5 defined according to step 1
The Inport And Outport Node of system, and the feature vector obtained in step 5 is inputted into small wave width learning system and is classified;It is instructing
According to the Sample Refreshment parameter for having label during white silk;In test and practical process, from the small wave width learning system obtained
The serial number of highest incentive degree node is judged in several nodes of output, then the serial number of the node is the serial number of classification movement;
The small wave width learning system specifically:
(1) different groups of characteristic nodes are mapped using input vector:
Wherein wi, aiAnd biWeight, transfer parameters and the scaling parameter respectively mapped, these parameters are in initialization
It is random to generate and updated using k-mean algorithm;φi() is i-th of wavelet basis function, and wherein n is that wavelet basis function is always a
Number;
(2) series connection each group characteristic node:
Zn=[Z1,Z2,......,Zn];
(3) characteristic node Z is utilizedn, map incremental nodes:
Hm=ξ (ZnWh+βh),
Wherein WhAnd βhThe respectively weight and threshold parameter of incremental nodes, this partial parameters generate at random in initialization
Just fixation does not need to update;ξ () is general excitation function, and sigmoid function may be used herein;
(4) in training process, the weighting parameter of output is obtained using pseudo- reciprocal value and ridge regression algorithm:
Wall=[Zn|Hm]+Y,
Wherein Y is the reference output of training set, and [Zn|Hm]+For [Zn|Hm] pseudo- reciprocal value;In test and practical process,
Output node can directly be mapped out
Step 7 can be known according to the corresponding serial number of each gesture is defined in step 1 by the serial number of classification results
Not Chu gesture type.
The present invention applies small wave width learning system in gesture classification recognizer and goes training and classify, and advantage exists
In combining width learning system, there is quicker and more efficient advantage compared to deep learning network;It uses
Excitation function of the wavelet basis function as characteristic layer, this is improved the nonlinear fitting ability of network.
The present invention can be more quickly completed the training and parameter of model compared to the algorithm based on conventional depth learning network
It determines, to improve working efficiency;Node is dynamically expanded to improve the discrimination of system, without building again completely
Vertical and training pattern.
Above-mentioned is the preferable embodiment of the present invention, but embodiments of the present invention are not limited by the foregoing content,
His any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the present invention, should be
The substitute mode of effect, is included within the scope of the present invention.
Claims (3)
1. a kind of sEMG gesture identification method based on small wave width learning system, which is characterized in that include the following steps:
Step 1, the quantity d of gesture motion type is defined according to corresponding identification scene, and distributes a sequence to each gesture
Number;
Step 2 acquires sEMG signal s using electromyographic signal collection equipment;
Step 3 is filtered noise reduction to sEMG original signal, while fertile hereby using Bart according to the frequency characteristic of sEMG signal
Filter removes the noise other than 10Hz~500Hz frequency band:
Wherein N is the order of filter, ωcFor cutoff frequency filter;
Step 4, using the active segment in Moving Window method detection sEMG signal;
Step 5 carries out feature extraction to the active segment that detection obtains;The feature of signal segment is calculated, this feature includes: averagely absolutely
To value q1, root mean square q2, median frequency q3, means frequency q4..., each feature for active segment of connecting is feature vector:
xk=[q1,q2,...,qn]T,
Wherein k is kth section active signal section;
Further, using the feature vector of each active segment as the input vector x of small wave width learning classifier system;
Step 6, the small wave width learning system of quantitative design of feature in the gesture quantity and step 5 defined according to step 1
Inport And Outport Node, and the feature vector obtained in step 5 is inputted into small wave width learning system and is classified;It was training
According to the Sample Refreshment parameter for having label in journey;In test and practical process, from the small wave width learning system output obtained
Several nodes in judge the serial number of highest incentive degree node, then the serial number of the node is the serial number of classification movement;
Step 7 can be identified according to the corresponding serial number of each gesture is defined in step 1 by the serial number of classification results
The type of gesture.
2. the sEMG gesture identification method according to claim 1 based on small wave width learning system, which is characterized in that institute
State the Moving Window method in step 4 specifically:
The signal data in a bit of time is extracted first and carries out the operation of integrated square, as shown in following formula:
Wherein s (t) is the electromyography signal data in window, SiIndicate tiThe energy value at the integral of time-ofday signals i.e. corresponding moment;Setting
One threshold value beta is to energy value SiIt makes a decision, works as tiThe energy value S at momentiGreater than the continuous n after threshold value beta and moving window1Secondary energy
Magnitude is all larger than threshold value, then it is assumed that moment tiAt the beginning of movement;On this basis, when from tjThe energy value S at momentj
Less than threshold value beta and there is continuous n2Secondary energy value is less than threshold value and then thinks tjTo act finish time, therefore, action signal Duan Ze are as follows:
Wherein k is the movement that kth section detected.
3. the sEMG gesture identification method according to claim 1 based on small wave width learning system, which is characterized in that institute
State small wave width learning system in step 6 specifically:
(1) different groups of characteristic nodes are mapped using input vector:
Wherein wi, aiAnd biWeight, transfer parameters and the scaling parameter respectively mapped, these parameters are random in initialization
It generates and k-mean algorithm is utilized to update;φi() is i-th of wavelet basis function, and wherein n is wavelet basis function total number;
(2) series connection each group characteristic node:
Zn=[Z1,Z2,......,Zn];
(3) characteristic node Z is utilizedn, map incremental nodes:
Hm=ξ (ZnWh+βh),
Wherein WhAnd βhThe respectively weight and threshold parameter of incremental nodes, this partial parameters generate just solid at random in initialization
Surely it does not need to update;ξ () is general excitation function, and sigmoid function may be used herein;
(4) in training process, the weighting parameter of output is obtained using pseudo- reciprocal value and ridge regression algorithm:
Wall=[Zn|Hm]+Y,
Wherein Y is the reference output of training set, and [Zn|Hm]+For [Zn|Hm] pseudo- reciprocal value;It, can be straight in test and practical process
It connects and maps out output node
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910548939.4A CN110389663B (en) | 2019-06-24 | 2019-06-24 | sEMG gesture recognition method based on wavelet width learning system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910548939.4A CN110389663B (en) | 2019-06-24 | 2019-06-24 | sEMG gesture recognition method based on wavelet width learning system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110389663A true CN110389663A (en) | 2019-10-29 |
CN110389663B CN110389663B (en) | 2023-05-23 |
Family
ID=68285868
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910548939.4A Active CN110389663B (en) | 2019-06-24 | 2019-06-24 | sEMG gesture recognition method based on wavelet width learning system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110389663B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110826625A (en) * | 2019-11-06 | 2020-02-21 | 南昌大学 | Finger gesture classification method based on surface electromyographic signals |
CN111160392A (en) * | 2019-12-03 | 2020-05-15 | 广东工业大学 | Hyperspectral classification method based on wavelet width learning system |
CN111695446A (en) * | 2020-05-26 | 2020-09-22 | 浙江工业大学 | Gesture recognition method integrating sEMG and AUS |
CN113657479A (en) * | 2021-08-12 | 2021-11-16 | 广东省人民医院 | Novel multi-scale depth-width combined pathological picture classification method, system and medium |
CN114098768A (en) * | 2021-11-25 | 2022-03-01 | 哈尔滨工业大学 | Cross-individual surface electromyographic signal gesture recognition method based on dynamic threshold and easy TL |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106236336A (en) * | 2016-08-15 | 2016-12-21 | 中国科学院重庆绿色智能技术研究院 | A kind of myoelectric limb gesture and dynamics control method |
US20170220923A1 (en) * | 2016-02-02 | 2017-08-03 | Samsung Electronics Co., Ltd. | Gesture classification apparatus and method using emg signal |
CN109062401A (en) * | 2018-07-11 | 2018-12-21 | 北京理工大学 | A kind of real-time gesture identifying system based on electromyography signal |
CN109077715A (en) * | 2018-09-03 | 2018-12-25 | 北京工业大学 | A kind of electrocardiosignal automatic classification method based on single lead |
CN109308521A (en) * | 2018-08-27 | 2019-02-05 | 广东工业大学 | A kind of quaternary SerComm degree study filtering method for eliminating Physiological tremor |
-
2019
- 2019-06-24 CN CN201910548939.4A patent/CN110389663B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170220923A1 (en) * | 2016-02-02 | 2017-08-03 | Samsung Electronics Co., Ltd. | Gesture classification apparatus and method using emg signal |
CN106236336A (en) * | 2016-08-15 | 2016-12-21 | 中国科学院重庆绿色智能技术研究院 | A kind of myoelectric limb gesture and dynamics control method |
CN109062401A (en) * | 2018-07-11 | 2018-12-21 | 北京理工大学 | A kind of real-time gesture identifying system based on electromyography signal |
CN109308521A (en) * | 2018-08-27 | 2019-02-05 | 广东工业大学 | A kind of quaternary SerComm degree study filtering method for eliminating Physiological tremor |
CN109077715A (en) * | 2018-09-03 | 2018-12-25 | 北京工业大学 | A kind of electrocardiosignal automatic classification method based on single lead |
Non-Patent Citations (1)
Title |
---|
李博等: "基于小波特征分析的手指动作识别研究", 《现代生物医学进展》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110826625A (en) * | 2019-11-06 | 2020-02-21 | 南昌大学 | Finger gesture classification method based on surface electromyographic signals |
CN110826625B (en) * | 2019-11-06 | 2022-04-12 | 南昌大学 | Finger gesture classification method based on surface electromyographic signals |
CN111160392A (en) * | 2019-12-03 | 2020-05-15 | 广东工业大学 | Hyperspectral classification method based on wavelet width learning system |
CN111695446A (en) * | 2020-05-26 | 2020-09-22 | 浙江工业大学 | Gesture recognition method integrating sEMG and AUS |
CN111695446B (en) * | 2020-05-26 | 2021-07-27 | 浙江工业大学 | Gesture recognition method integrating sEMG and AUS |
CN113657479A (en) * | 2021-08-12 | 2021-11-16 | 广东省人民医院 | Novel multi-scale depth-width combined pathological picture classification method, system and medium |
CN113657479B (en) * | 2021-08-12 | 2022-12-06 | 广东省人民医院 | Novel multi-scale depth-width combined pathological picture classification method, system and medium |
CN114098768A (en) * | 2021-11-25 | 2022-03-01 | 哈尔滨工业大学 | Cross-individual surface electromyographic signal gesture recognition method based on dynamic threshold and easy TL |
CN114098768B (en) * | 2021-11-25 | 2024-05-03 | 哈尔滨工业大学 | Cross-individual surface electromyographic signal gesture recognition method based on dynamic threshold and EasyTL |
Also Published As
Publication number | Publication date |
---|---|
CN110389663B (en) | 2023-05-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110389663A (en) | A kind of sEMG gesture identification method based on small wave width learning system | |
CN106169081B (en) | A kind of image classification and processing method based on different illumination | |
CN103996018B (en) | Face identification method based on 4DLBP | |
CN105760821B (en) | The face identification method of the grouped accumulation rarefaction representation based on nuclear space | |
CN101763502B (en) | High-efficiency method and system for sensitive image detection | |
CN105844285A (en) | Cucumber disease identification method and apparatus based on image information | |
CN101859377A (en) | Electromyographic signal classification method based on multi-kernel support vector machine | |
CN110827260B (en) | Cloth defect classification method based on LBP characteristics and convolutional neural network | |
CN105740887A (en) | Electroencephalogram feature classification method based on PSO-SVM (Particle Swarm Optimization-Support Vector Machine) | |
CN101526994B (en) | Fingerprint image segmentation method irrelevant to collecting device | |
CN106709528A (en) | Method and device of vehicle reidentification based on multiple objective function deep learning | |
CN103593652B (en) | A kind of cucumber disease recognition methods processed based on cucumber leaves symptomatic picture | |
CN109344699A (en) | Winter jujube disease recognition method based on depth of seam division convolutional neural networks | |
CN105405136A (en) | Self-adaptive spinal CT image segmentation method based on particle swarm optimization | |
CN106647272A (en) | Robot route planning method by employing improved convolutional neural network based on K mean value | |
CN110558975B (en) | Electrocardiosignal classification method and system | |
CN105069745A (en) | face-changing system based on common image sensor and enhanced augmented reality technology and method | |
CN107729890A (en) | Face identification method based on LBP and deep learning | |
CN109730818A (en) | A kind of prosthetic hand control method based on deep learning | |
CN107025446A (en) | A kind of vibration signal combines noise-reduction method | |
CN105550705A (en) | Electroencephalogram (EEG) signal recognition method based on improved self-training learning | |
CN109145135A (en) | A kind of human face portrait aging method based on principal component analysis | |
CN108509861A (en) | A kind of method for tracking target and device combined based on sample learning and target detection | |
CN104573701B (en) | A kind of automatic testing method of Tassel of Corn | |
CN111461000A (en) | Intelligent office garbage classification method based on CNN and wavelet analysis |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |