CN107157450A - Quantitative estimation method and system are carried out for the hand exercise ability to patient Parkinson - Google Patents
Quantitative estimation method and system are carried out for the hand exercise ability to patient Parkinson Download PDFInfo
- Publication number
- CN107157450A CN107157450A CN201710463520.XA CN201710463520A CN107157450A CN 107157450 A CN107157450 A CN 107157450A CN 201710463520 A CN201710463520 A CN 201710463520A CN 107157450 A CN107157450 A CN 107157450A
- Authority
- CN
- China
- Prior art keywords
- required movement
- feature
- training dataset
- mrow
- grader
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/48—Other medical applications
- A61B5/4848—Monitoring or testing the effects of treatment, e.g. of medication
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
- A61B5/1124—Determining motor skills
- A61B5/1125—Grasping motions of hands
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/389—Electromyography [EMG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/40—Detecting, measuring or recording for evaluating the nervous system
- A61B5/4076—Diagnosing or monitoring particular conditions of the nervous system
- A61B5/4082—Diagnosing or monitoring movement diseases, e.g. Parkinson, Huntington or Tourette
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Veterinary Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Artificial Intelligence (AREA)
- Pathology (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physiology (AREA)
- Neurology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Neurosurgery (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Developmental Disabilities (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Mathematical Physics (AREA)
- Fuzzy Systems (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Dentistry (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention provides carry out quantitative estimation method and system for the hand exercise ability to patient Parkinson, using hand surface electromyogram signal of the wearable myoelectric sensor capture detected person when performing required movement, quantitative evaluation is carried out to the hand exercise ability of detected person to extract temporal signatures and frequency domain character and the feature related with the performance of required movement from surface electromyogram signal based on the locomitivity grader corresponding to the required movement trained.Hand exercise ability that can be more objectively and accurately to patient Parkinson by this method and system carries out grade assessment.
Description
Technical field
Commented the present invention relates to the quantization of the qualitative assessment of locomitivity, more particularly to the hand exercise ability of patient Parkinson
Estimate method and system.
Background technology
The motor symptoms of Parkinson's (Parkinson ' s Disease, PD) patient is mainly shown as bradykinesia, static
Property is trembled with muscular rigidity etc..The detection and assessment of the current locomitivity for PD patient depend primarily on international movement association
The PD measuring scales (UnifiedParkinson's Disease Rating Scale, UPDRS) that can be provided are commented to carry out grade
It is fixed, for example, the hand exercise ability of generally patient Parkinson is divided into 0-4 grades, 0 grade of hand exercise energy equivalent to normal person
Power, series is higher, and hand exercise ability is poorer.However, this evaluation mode is generally by scoring doctor's operating experience and assessment
When detected person state and mood influence, therefore its assess result it is still not objective and accurate enough.
The content of the invention
Therefore, it is an object of the invention to the defect for overcoming above-mentioned prior art, there is provided a kind of surface flesh of utilization hand
Quantitative estimation method and system of the electric signal to the hand exercise ability of patient Parkinson.
The purpose of the present invention is achieved through the following technical solutions：
On the one hand, it is used to carry out quantitative evaluation side to the hand exercise ability of patient Parkinson the invention provides a kind of
Method, including：
Step 1, via be placed on detected person's hand myoelectric sensor gather detected person perform required movement when
Surface electromyogram signal；
Step 2, using the good patient's Parkinson locomitivity grader corresponding with the required movement of training in advance come
Surface electromyogram signal according to being gathered judges the grade belonging to the hand exercise ability of detected person；
Wherein it is used to train the feature of patient's Parkinson locomitivity grader corresponding with required movement to include from surface
Temporal signatures and frequency domain character are extracted in electromyographic signal and based on the complete with the required movement of surface electromyogram signal acquisition
Into the related feature of situation.
In the above method, the feature related to the performance of the required movement may include it is following at least one：
The time used in maximum electromyographic signal, completion required movement during completion required movement.
In the above method, the required movement may include it is following at least one：Clench fist, to referring to.
In the above method, the step for training patient's Parkinson locomitivity grader corresponding with required movement may also include
Suddenly, it includes：
Worn when a) receiving from multiple different degrees of patients Parkinson and normal person's execution required movement by it
Myoelectric sensor collection multiple surface electromyogram signals be used as training dataset；
B) extracted from each surface electromyogram signal temporal signatures and frequency domain character and based on surface electromyogram signal obtain with
The related feature of the performance of the required movement；
C) select to be used to train the locomitivity grader for the information gain of training dataset based on each feature
Feature, wherein each feature to the information gain of training dataset for training dataset empirical entropy with given this feature
Under the conditions of training dataset empirical condition entropy between difference；
D) the locomitivity grader is trained based on selected feature.
In the above method, when the required movement is to referring to, the motion can be used as using the minimum Optimized model of sequence
Capability Categories device；When the required movement is to clench fist, the locomitivity grader can be used as using J48 graders.
In the above method, the empirical entropy of training dataset represents the uncertainty classified to training dataset, can be with
Equation below is calculated：
Wherein D represents training dataset, and H (D) represents training dataset D empirical entropy, and n represents that training dataset D is divided into
For n classes, piExpression divides the data into the probability of the i-th class.
In the above method, the empirical condition entropy of training dataset is represented in the given spy under conditions of some feature is given
The uncertainty classified under conditions of levying to training dataset, can be calculated with equation below：
Wherein A represents some feature, the empirical entropy that H (D | A) is the training dataset D under conditions of given feature A, pi(D
| A) represent to divide the data into the probability of the i-th class in the case of known features A.
Another aspect, is used to carry out quantitative evaluation system to the hand exercise ability of patient Parkinson the invention provides a kind of
System, including：
Harvester, specified move is performed for gathering detected person via the myoelectric sensor for being placed on detected person's hand
Surface electromyogram signal when making；
Detection means, for utilizing the good patient's Parkinson locomitivity corresponding with the required movement of training in advance point
Class device judges the grade belonging to the hand exercise ability of detected person according to the surface electromyogram signal gathered；
Wherein it is used to train the feature of patient's Parkinson locomitivity grader corresponding with required movement to include from surface
Temporal signatures and frequency domain character are extracted in electromyographic signal and based on the complete with the required movement of surface electromyogram signal acquisition
Into the related feature of situation.
In said system, the feature related to the performance of the required movement may include it is following at least one：
The time used in maximum electromyographic signal, completion required movement during completion required movement.
In said system, the required movement may include it is following at least one：Clench fist, to referring to.
Said system may also include the instruction for training patient's Parkinson locomitivity grader corresponding with required movement
Practice device, it is configured as：
Worn when a) receiving from multiple different degrees of patients Parkinson and normal person's execution required movement by it
Myoelectric sensor collection multiple surface electromyogram signals be used as training dataset；
B) extracted from each surface electromyogram signal temporal signatures and frequency domain character and based on surface electromyogram signal obtain with
The related feature of the performance of the required movement；
C) select to be used to train the locomitivity grader for the information gain of training dataset based on each feature
Feature, wherein each feature to the information gain of training dataset for training dataset empirical entropy with given this feature
Under the conditions of training dataset empirical condition entropy between difference；
D) the locomitivity grader is trained based on selected feature.
In said system, when the required movement is to referring to, the motion can be used as using the minimum Optimized model of sequence
Capability Categories device；When the required movement is to clench fist, the locomitivity grader can be used as using J48 graders.
Compared with prior art, the advantage of the invention is that：
Detected person's hand surface electromyogram signal is captured using wearable myoelectric sensor, with from surface electromyogram signal
Extract temporal signatures and frequency domain character and the feature related to the performance of required movement is transported the hand of detected person
Kinetic force carries out quantitative evaluation, and hand exercise ability that can be more objectively and accurately to patient Parkinson carries out grade assessment.
Brief description of the drawings
Embodiments of the present invention is further illustrated referring to the drawings, wherein：
Fig. 1 is to be used for the hand exercise ability progress quantitative estimation method to patient Parkinson according to the embodiment of the present invention
Schematic flow sheet；
Fig. 2 (a) is the accuracy of identification comparison schematic diagram of each grader corresponding with " to referring to " action；
Fig. 2 (b) is the accuracy of identification comparison schematic diagram of each grader corresponding with " clenching fist " action；
Fig. 3 (a) is to be known for " to referring to " action using the SMO graders of the signal identification features training of surface electromyogram signal
Other performance schematic diagram；
Fig. 3 (b) is to act to know using the J48 graders of the signal identification features training of surface electromyogram signal for " clenching fist "
Other performance schematic diagram；
Fig. 4 (a) is that the grader recognition performance corresponding with " to referring to " action trained using the method according to the invention is shown
It is intended to；
Fig. 4 (b) is that the grader recognition performance corresponding with " clenching fist " action trained using the method according to the invention is shown
It is intended to；
Fig. 5 (a) is to be illustrated using being contrasted with the recognition result of " to referring to " corresponding grader of action for different characteristic training
Figure；
Fig. 5 (b) is that trained using different characteristic contrasted with the corresponding grader of " clenching fist " action recognition result is illustrated
Figure.
Embodiment
In order that the purpose of the present invention, technical scheme and advantage are more clearly understood, pass through below in conjunction with accompanying drawing specific real
Applying example, the present invention is described in more detail.It should be appreciated that specific embodiment described herein is only to explain the present invention, and
It is not used in the restriction present invention.
" to referring to " and " clenching fist " gesture are two kinds of conventional worlds for assessing the hand exercise ability of patient Parkinson
Standard index.Hereinafter, it will in conjunction with specific embodiments illustrate how to assess Parkinson's by taking both required movements as an example
The hand exercise ability of people.It should be understood that embodiments of the invention are for specific athletic performance and are not limited, the present invention
Principle can also be applied to other action or other positions locomitivity assess.
Fig. 1, which gives, according to an embodiment of the invention to be used to measure the hand exercise ability of patient Parkinson
Change appraisal procedure, it includes performing via the myoelectric sensor collection detected person for the region of interest for being placed on detected person's hand
Surface electromyogram signal (step 1) during required movement；Utilize the good Parkinson's corresponding with the required movement of training in advance
People's locomitivity grader come judged according to the surface electromyogram signal gathered belonging to the hand exercise ability of detected person etc.
Level (step 2).Patient's Parkinson locomitivity grader wherein corresponding with required movement is in substantial amounts of different degrees of handkerchief
The gloomy patient of gold and some normal persons perform what is be trained on the basis of the sample data extracted during the required movement.
More specifically, in step 1, myoelectric sensor is placed on to the specified location of detected person, the correlation such as arm
Position, then allows detected person to perform required movement, for example, clench fist or to referring to.Here it is possible to using people's body surface can be collected
Any kind of myoelectric sensor of facial muscle electric signal.Preferably, wearable myoelectric apparatus can be used, for example, is hereinafter used
Surface electromyogram signal during hand exercise is detected by the armlet MYO of Canadian venture company ThalmicLabs products, it is a kind of
Gesture control armlet, can carry out wireless connection and data transfer by bluetooth and other electronic products.Detected person is holding
Fist and when being acted to finger, the muscle carry out activity mainly above forearm produces corresponding electromyographic signal, it is possible to by MYO
Above the forearm for being placed on detected person.The surface electromyogram signal collected is passed through wired or wireless transmission by myoelectric sensor
Mode is sent to corresponding signal analysis and processing device, for example, the mobile terminal of such as flat board or mobile phone etc, desktop computer
And any other computing device of surface electromyogram signal can be handled.
Preferably, being responsible for the computing device of processing surface electromyogram signal can also be carried out to the surface electromyogram signal received
A certain degree of pretreatment, for example, filter out noise or carry out dimension-reduction treatment, such as to the letter from multiple myoelectric sensors
Number synthesized, to simplify computation complexity as far as possible, reduce the consumption to computing resource.By taking MYO as an example, it is by eight sensors
Chip is constituted, and can collect the electromyographic signal of one week common octuple of patient's forearm, a is used respectively1,a2,a3,a4,a5,a6,a7,a8Come
Represent.When detected person does " clenching fist " and " to referring to ", the muscle on forearm surface can be active.For example when patient clenches fist
When, the forearm muscle of one week in tight state, that is, state of activation, causes eight myoelectric sensors on myoelectric apparatus simultaneously
Electromyographic signal increase simultaneously.When patient's arm opens, the muscle of one week is in relaxation state simultaneously on forearm, or is non-
State of activation, causes the electromyographic signal of eight myoelectric sensors while reducing or disappearing.The octuple electromyographic signal collected
The correlation having, it is possible thereby to synthesized the electromyographic signal gathered using the method for resultant acceleration, such as it is public
Formula (1)So as to the electromyographic signal sample synthesized.
Using synthesis electromyographic signal while computation complexity is reduced, wearer can also be allowed to ignore the direction for wearing myoelectric apparatus, can
Arbitrarily to wear the myoelectric apparatus of this eight myoelectricity paster compositions.It should be understood that pretreatment carries out being preferred mode rather than right above
This carries out any limitation, and the surface electromyogram signal gathered using single myoelectric sensor can also realize the amount of hand exercise ability
Change and assess.No longer refered in particular to during the surface electromyogram signal hereinafter mentioned the collection of single myoelectric sensor electromyographic signal or
Electromyographic signal after synthesis, because the processing mode of the two is similar.
With continued reference to Fig. 1, in step 2, after surface electromyogram signal when performing required movement is obtained, carry therefrom
The feature required for patient's Parkinson locomitivity grader corresponding with the required movement is taken, then the feature extracted is put forward
The good grader of training in advance is supplied to judge the grade belonging to the hand exercise ability of detected person.Specify dynamic for each
Make, there is patient's Parkinson locomitivity grader corresponding thereto that training in advance is good, will be with holding for method is described below
Fist acts corresponding grader and is referred to as grader of clenching fist, and will referred to as classify with to the corresponding grader of finger action to referring to
Device.The training method of the two graders is similar, therefore, and the training method to grader is hereafter only carried out by taking grader of clenching fist as an example
It is illustrated.
Substantial amounts of different degrees of patient Parkinson and some normal persons are allowed to wear myoelectric sensor simultaneously first in training
Execution is clenched fist action, so as to collect substantial amounts of surface electromyogram signal as basic training dataset.When training grader
Key is to determine which feature using data to be trained, and these features can directly affect the accuracy of classification results.It is logical
Often, time and frequency domain analysis can be carried out to surface electromyogram signal, therefrom extracts multiple temporal signatures and frequency domain character to train
Grader.The characteristics of these time domains and frequency domain character reflect electromyographic signal waveform itself.In a preferred embodiment of the invention,
In addition to using temporal signatures and frequency domain character, the feature related to the performance of required movement is additionally used.This is to examine
Considering different degrees of PD patient time, the strength used when completing some required movement etc. has larger difference, therefore, with finger
Surely the feature of the performance correlation acted can also have preferable distinction for hand exercise ability.In order to improve grader
The accuracy of itself, is selected by way of based on information gain from above-mentioned multiple features in an embodiment of the present invention
Optimal some features train grader.The process of training grader mainly includes：
(1) temporal signatures and frequency domain character are extracted from surface electromyogram signal
Surface electromyogram signal (surface electromyogram signal, SEMS) is passed through from human skin
The bioelectrical signals provided when sensor or electrode record neuron-muscular activity, belong to non-invasive.The amplitude of surface electromyogram signal
With randomness, quasi- gauss of distribution function can be typically expressed as.Usual digital assay is the main of processing surface electromyogram signal
Means, including time-domain analysis and frequency-domain analysis.Time-domain analysis is the function that surface electromyogram signal is regarded as to the time, by analyzing
To some statistical natures of electromyographic signal, for example, common temporal signatures include average of the electromyographic signal in time domain, a window
Intraoral maximum, variance, standard deviation, mode, zero passage number of times, mean-square value, mistake average rate etc..Frequency-domain analysis is by Fourier
Become time-domain signal of changing commanders and be converted to frequency-region signal, commonly using frequency domain character includes peak value, frequency component, DC component, energy, shape
Average, shape criteria are poor, the shape degree of bias, shape kurtosis, amplitude equalizing value, poor amplitude criteria, the amplitude degree of bias, amplitude kurtosis, intermediate value
Frequency, means frequency, frequency range, highest peak frequencies, highest crest amplitude etc..Above-mentioned time domain and frequency domain character can also
It is referred to as signal identification feature.
(2) feature related to the performance of required movement is obtained based on surface electromyogram signal
The feature (may be simply referred to as motion characteristic) related to the performance of required movement is including completing during required movement
Time used in maximum electromyographic signal, completion required movement etc..With the exacerbation of PD conditions of patients, Parkinson's motor symptoms meeting
There is pause and slow, this will cause the decline of electromyographic signal intensity.In addition, when PD motor symptomses are relatively serious, completing dynamic
Making the time used can be elongated therewith.Therefore, the feature related from the performance of required movement can also efficiently differentiate different journeys
The hand exercise ability of the PD patient of degree.
(3) validity feature is selected to train from above-mentioned multiple signal identification features and motion characteristic based on information gain
Grader.
First, the information gain of each feature is calculated.Information gain represents to learn feature X information and causes class Y letter
The uncertain reduction degree of breath.Feature A is designated as g (D, A) to training dataset D information gain, and it is training data set D
Empirical entropy or comentropy H (D) and under feature A specified criterias between training data set D empirical condition entropy H (D | A)
Difference, i.e.,：
G (D, A)=H (D)-H (D | A)
The uncertainty that wherein empirical entropy H (D) expressions are classified to data set D, and empirical condition entropy H (D | A) represent
The uncertainty classified under conditions of feature A is given to data set D, difference therebetween is information gain, is represented
Due to feature A so that the degree of the uncertain reduction to data set D classification, it is clear that for training dataset D, letter
Cease gain and depend on feature, different features often have different information gains, and the big feature of information gain has stronger
Classification capacity.Empirical entropy H (D) can for example be calculated by following formula：
Wherein n represents that training dataset D is divided into n classes, piExpression divides the data into the probability of the i-th class, as known to above formula
Entropy H is bigger, and the uncertainty of classification is bigger.Empirical condition entropy H (D | A) it can for example be calculated by following formula：
Wherein n represents that training dataset D is divided into n classes, pi(D | A) represent in the case of known features A data point
For the probability of the i-th class, H (D | A) be characterized in feature A it is given under conditions of the degree of uncertainty classified to data set D.
After the information gain of each feature of PD motor symptomses can be calculated according to above-mentioned formula, compare each feature
Information gain, selects the larger feature of those information gains as training grader, from the surface of training data concentration
These features are extracted in electromyographic signal as sample data to train grader.
In order to further illustrate the effect of the method according to above-described embodiment, inventor has also carried out following experiment.
First, grader is trained using only the time domain and frequency domain character extracted from surface electromyogram signal.For example distinguish
Use support vector machines (Support Vector Machine), sequence minimum optimization SMO (Sequential minimal
Optimization) model, such as J48 decision-tree model and random forest RF (RandomForest) models to " clenching fist " and
The training data of " to referring to " action is handled, and experimental result such as Fig. 2 (a) and 2 (b) are shown.For " clenching fist " action, use
The accuracy of identification highest of J48 graders, can reach 86.2%；For " to refer to " action, using SMO graders accuracy of identification most
Height, can reach 63.7%.Wherein, the precision of its affiliated grade (i.e. 0-4 grades) is recognized using SMO graders for " to referring to " action
(precision), shown in recall rate (Recall) and F-score result such as Fig. 3 (a), for each grade, these three indexs
From left to right it is arranged in order in figure." clenching fist " is acted, the essence of its affiliated grade (i.e. 0-4 grades) is recognized using J48 graders
Shown in result such as Fig. 3 (b) of degree, recall rate and F-score.By observing electromyographic signal intensity, the electromyographic signal to finger is found
It is weak compared with the electromyographic signal clenched fist, because the muscle activation degree to finger is weaker compared with the muscle activation degree clenched fist, and myoelectricity
It is more difficult when signal is relatively weak to be classified, thus it is unobvious when relatively being clenched fist to the differentiation of the electromyographic signal of finger, to finger during identification
The height that precision is not clenched fist.When classification grade is 0 and 1, error rate is higher, that is to say, that normal person to referring to grade and action
Slight slack-off patient is not easily distinguishable to referring to grade, this be due to normal person after substantial amounts of work work, can also produce action
Slight slack-off symptom.When classification grade is 2 and 3, precision is relatively low, such as shown in Fig. 3 (a).Because patient is being done to referring to
During test, grade is weaker for the muscle activation degree of 2 and 3 patient, causes electromyographic signal intensity less than normal, and the little flesh of intensity
When slight change occurs for electric signal, the result that classification is obscured is easily caused.The above factor is caused to referring to classification error rate
High main cause.
Then, acted for " clenching fist ", using J48 graders, for " to referring to " action, using SMO graders, but only
Classification is trained using maximum electromyographic signal when completing required movement, these motion characteristics of time completed used in required movement
Device.When using J48 graders, the nicety of grading for grader of clenching fist reaches 70.6%, when using SMO graders, divides referring to
The accuracy of identification of class device reaches 72.5%.
Finally, N before being selected by way of above-mentioned calculating information gain from above-mentioned signal identification feature and motion characteristic
Individual validity feature trains grader, using sorter model same as above.Such as when using J48 graders, clench fist point
The precision of class device reaches 90.8%.When using SMO graders, 82.3% is reached to the precision for referring to grader.It is such to specify
Classification of motion device is shown for accuracy of identification, recall rate and F-score such as Fig. 4 (a)-(b) of hand ability 0-4 grades, wherein
For each grade, these three indexs are from left to right arranged in order in figure.Shown in Fig. 5 (a)-(b) is respectively to be known with signal
The experimental result comparison diagram of other feature, motion characteristic and fusion feature.As shown in figure 5, with using only signal identification feature or
It is only applicable to motion characteristic to compare, point after validity feature (being referred to as fusion feature in figure) is selected by the way of information gain
Class precision is significantly improved.
Although the present invention be described by means of preferred embodiments, but the present invention be not limited to it is described here
Embodiment, without departing from the present invention also include made various changes and change.
Claims (10)
1. a kind of be used to carry out quantitative estimation method to the hand exercise ability of patient Parkinson, methods described includes：
Step 1, surface when detected person performs required movement is gathered via the myoelectric sensor for being placed on detected person's hand
Electromyographic signal；
Step 2, using the good patient's Parkinson locomitivity grader corresponding with the required movement of training in advance come basis
The surface electromyogram signal gathered judges the grade belonging to the hand exercise ability of detected person；
Wherein it is used to train the feature of patient's Parkinson locomitivity grader corresponding with required movement to include from surface myoelectric
Temporal signatures and frequency domain character and the completion feelings with the required movement obtained based on surface electromyogram signal are extracted in signal
The related feature of condition.
2. according to the method described in claim 1, wherein the feature related to the performance of the required movement is including following
In at least one：The time used in maximum electromyographic signal, completion required movement during completion required movement.
3. according to the method described in claim 1, wherein the required movement include it is following at least one：Clench fist, to referring to.
4. the method according to any one of claim 1-3, in addition to train patient Parkinson corresponding with required movement
The step of locomitivity grader, it includes：
A) flesh worn during from multiple different degrees of patients Parkinson and normal person's execution required movement by it is received
Multiple surface electromyogram signals of electric transducer collection are used as training dataset；
B) extracted from each surface electromyogram signal temporal signatures and frequency domain character and based on surface electromyogram signal obtain with it is described
The related feature of the performance of required movement；
C) spy for training the locomitivity grader is selected for the information gain of training dataset based on each feature
Levy, wherein each feature is to the information gain of training dataset empirical entropy and the condition in given this feature for training dataset
Difference between the empirical condition entropy of lower training dataset；
D) the locomitivity grader is trained based on selected feature.
5. method according to claim 4, wherein when the required movement is to referring to, using the minimum Optimized model of sequence
It is used as the locomitivity grader；When the required movement is to clench fist, the locomitivity point is used as using J48 graders
Class device.
6. method according to claim 4, wherein, the empirical entropy of training dataset represents to classify to training dataset
Uncertainty, calculated with equation below：
<mrow>
<mi>H</mi>
<mrow>
<mo>(</mo>
<mi>D</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mo>-</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msub>
<mi>p</mi>
<mi>i</mi>
</msub>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mi> </mi>
<msub>
<mi>p</mi>
<mi>i</mi>
</msub>
</mrow>
Wherein D represents training dataset, and H (D) represents training dataset D empirical entropy, and n represents that training dataset D is divided into n
Class, piExpression divides the data into the probability of the i-th class.
7. method according to claim 6, wherein under conditions of some feature is given training dataset empirical condition
Entropy represents the uncertainty classified under conditions of given this feature to training dataset, carried out by equation below in terms of
Calculate：
<mrow>
<mi>H</mi>
<mrow>
<mo>(</mo>
<mi>D</mi>
<mo>|</mo>
<mi>A</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mo>-</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msub>
<mi>p</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>D</mi>
<mo>|</mo>
<mi>A</mi>
<mo>)</mo>
</mrow>
<msub>
<mi>logp</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>D</mi>
<mo>|</mo>
<mi>A</mi>
<mo>)</mo>
</mrow>
</mrow>
Wherein A represents some feature, the empirical entropy that H (D | A) is the training dataset D under conditions of given feature A, pi(D | A) table
Show the probability that the i-th class is divided the data into the case of known features A.
8. a kind of be used to carry out quantitative evaluation system to the hand exercise ability of patient Parkinson, the system includes：
Harvester, when performing required movement for gathering detected person via the myoelectric sensor for being placed on detected person's hand
Surface electromyogram signal；
Detection means, for utilizing the good patient's Parkinson locomitivity grader corresponding with the required movement of training in advance
To judge the grade belonging to the hand exercise ability of detected person according to the surface electromyogram signal gathered；
Wherein it is used to train the feature of patient's Parkinson locomitivity grader corresponding with required movement to include from surface myoelectric
Temporal signatures and frequency domain character and the completion feelings with the required movement obtained based on surface electromyogram signal are extracted in signal
The related feature of condition.
9. system according to claim 1, wherein the feature related to the performance of the required movement is including following
In at least one：The time used in maximum electromyographic signal, completion required movement during completion required movement.
10. method according to claim 1 or 2, in addition to for training patient's Parkinson fortune corresponding with required movement
The trainer of kinetic force grader, it is configured as：
Passed when a) performing the required movement from multiple different degrees of patients Parkinson and normal person by its myoelectricity worn
Multiple surface electromyogram signals of sensor collection are used as training dataset；
B) extracted from each surface electromyogram signal temporal signatures and frequency domain character and based on surface electromyogram signal obtain with it is described
The related feature of the performance of required movement；
C) spy for training the locomitivity grader is selected for the information gain of training dataset based on each feature
Levy, wherein each feature is to the information gain of training dataset empirical entropy and the condition in given this feature for training dataset
Difference between the empirical condition entropy of lower training dataset；
D) the locomitivity grader is trained based on selected feature.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710463520.XA CN107157450B (en) | 2017-06-19 | 2017-06-19 | Quantitative assessment method and system for hand motion ability of Parkinson's disease people |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710463520.XA CN107157450B (en) | 2017-06-19 | 2017-06-19 | Quantitative assessment method and system for hand motion ability of Parkinson's disease people |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107157450A true CN107157450A (en) | 2017-09-15 |
| CN107157450B CN107157450B (en) | 2020-03-31 |
Family
ID=59818815
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710463520.XA Active CN107157450B (en) | 2017-06-19 | 2017-06-19 | Quantitative assessment method and system for hand motion ability of Parkinson's disease people |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107157450B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108309295A (en) * | 2018-02-11 | 2018-07-24 | 宁波工程学院 | A kind of arm muscular strength assessment method |
| CN108670713A (en) * | 2018-04-23 | 2018-10-19 | 中国科学院深圳先进技术研究院 | A kind of exoskeleton robot |
| CN108742612A (en) * | 2018-05-24 | 2018-11-06 | 杭州航弈生物科技有限责任公司 | DBS validation checking equipment based on myoelectricity marker |
| CN109009148A (en) * | 2018-08-24 | 2018-12-18 | 广东工业大学 | A kind of gait function appraisal procedure |
| CN110516762A (en) * | 2019-10-10 | 2019-11-29 | 深圳大学 | A kind of muscular states quantization assessment method, apparatus, storage medium and intelligent terminal |
| CN110547807A (en) * | 2019-09-17 | 2019-12-10 | 深圳市赛梅斯凯科技有限公司 | driving behavior analysis method, device, equipment and computer readable storage medium |
| CN111210912A (en) * | 2020-01-14 | 2020-05-29 | 上海恩睦信息科技有限公司 | Parkinson prediction method and device |
| CN111415746A (en) * | 2020-04-22 | 2020-07-14 | 上海邦邦机器人有限公司 | Physical function evaluation model generation method, physical function evaluation method, and physical function evaluation apparatus |
| CN111528842A (en) * | 2020-05-26 | 2020-08-14 | 复嶂环洲生物科技(上海)有限公司 | Quantitative assessment method for Parkinson disease symptoms based on physiological and behavioral indexes |
| CN112826504A (en) * | 2021-01-07 | 2021-05-25 | 中新国际联合研究院 | Games-based Parkinson symptom grade assessment method and device |
| CN117643456A (en) * | 2024-01-29 | 2024-03-05 | 北京航空航天大学 | Auxiliary evaluation system, method and storage medium for parkinsonism |
| CN117643456B (en) * | 2024-01-29 | 2024-05-24 | 北京航空航天大学 | Auxiliary evaluation system, method and storage medium for parkinsonism |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050234309A1 (en) * | 2004-01-07 | 2005-10-20 | David Klapper | Method and apparatus for classification of movement states in Parkinson's disease |
| US20100176952A1 (en) * | 2008-12-04 | 2010-07-15 | The Regents Of The University Of California | System for detection of body motion |
| CN104207782A (en) * | 2014-08-22 | 2014-12-17 | 南昌大学 | Hand motion function rating device for Parkinson's disease patient |
| CN104398263A (en) * | 2014-12-25 | 2015-03-11 | 中国科学院合肥物质科学研究院 | Method for quantitatively evaluating symptoms of tremor of patient with Parkinson's disease according to approximate entropy and cross approximate entropy |
| CN104434129A (en) * | 2014-12-25 | 2015-03-25 | 中国科学院合肥物质科学研究院 | Quantization evaluating device and method for dyskinesia symptoms of Parkinson and related extrapyramidal diseases |
| CN104522949A (en) * | 2015-01-15 | 2015-04-22 | 中国科学院苏州生物医学工程技术研究所 | Smart wristband for quantitatively evaluating motion function of Parkinson patient |
| CN105426842A (en) * | 2015-11-19 | 2016-03-23 | 浙江大学 | Support vector machine based surface electromyogram signal multi-hand action identification method |
| CN105426696A (en) * | 2015-12-24 | 2016-03-23 | 中国科学院苏州生物医学工程技术研究所 | Multi-node quantitative assessment system and method for symptoms of Parkinson's disease |
| CN105550583A (en) * | 2015-12-22 | 2016-05-04 | 电子科技大学 | Random forest classification method based detection method for malicious application in Android platform |
| CN105930663A (en) * | 2016-04-26 | 2016-09-07 | 北京科技大学 | Parkinson's disease early diagnosis method |
-
2017
- 2017-06-19 CN CN201710463520.XA patent/CN107157450B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050234309A1 (en) * | 2004-01-07 | 2005-10-20 | David Klapper | Method and apparatus for classification of movement states in Parkinson's disease |
| US20100176952A1 (en) * | 2008-12-04 | 2010-07-15 | The Regents Of The University Of California | System for detection of body motion |
| CN104207782A (en) * | 2014-08-22 | 2014-12-17 | 南昌大学 | Hand motion function rating device for Parkinson's disease patient |
| CN104398263A (en) * | 2014-12-25 | 2015-03-11 | 中国科学院合肥物质科学研究院 | Method for quantitatively evaluating symptoms of tremor of patient with Parkinson's disease according to approximate entropy and cross approximate entropy |
| CN104434129A (en) * | 2014-12-25 | 2015-03-25 | 中国科学院合肥物质科学研究院 | Quantization evaluating device and method for dyskinesia symptoms of Parkinson and related extrapyramidal diseases |
| CN104522949A (en) * | 2015-01-15 | 2015-04-22 | 中国科学院苏州生物医学工程技术研究所 | Smart wristband for quantitatively evaluating motion function of Parkinson patient |
| CN105426842A (en) * | 2015-11-19 | 2016-03-23 | 浙江大学 | Support vector machine based surface electromyogram signal multi-hand action identification method |
| CN105550583A (en) * | 2015-12-22 | 2016-05-04 | 电子科技大学 | Random forest classification method based detection method for malicious application in Android platform |
| CN105426696A (en) * | 2015-12-24 | 2016-03-23 | 中国科学院苏州生物医学工程技术研究所 | Multi-node quantitative assessment system and method for symptoms of Parkinson's disease |
| CN105930663A (en) * | 2016-04-26 | 2016-09-07 | 北京科技大学 | Parkinson's disease early diagnosis method |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108309295A (en) * | 2018-02-11 | 2018-07-24 | 宁波工程学院 | A kind of arm muscular strength assessment method |
| CN108670713A (en) * | 2018-04-23 | 2018-10-19 | 中国科学院深圳先进技术研究院 | A kind of exoskeleton robot |
| CN108742612B (en) * | 2018-05-24 | 2023-12-12 | 王开亮 | DBS effectiveness detection equipment based on myoelectric marker |
| CN108742612A (en) * | 2018-05-24 | 2018-11-06 | 杭州航弈生物科技有限责任公司 | DBS validation checking equipment based on myoelectricity marker |
| CN109009148A (en) * | 2018-08-24 | 2018-12-18 | 广东工业大学 | A kind of gait function appraisal procedure |
| CN110547807A (en) * | 2019-09-17 | 2019-12-10 | 深圳市赛梅斯凯科技有限公司 | driving behavior analysis method, device, equipment and computer readable storage medium |
| CN110516762A (en) * | 2019-10-10 | 2019-11-29 | 深圳大学 | A kind of muscular states quantization assessment method, apparatus, storage medium and intelligent terminal |
| CN110516762B (en) * | 2019-10-10 | 2022-11-15 | 深圳大学 | Muscle state quantitative evaluation method and device, storage medium and intelligent terminal |
| CN111210912A (en) * | 2020-01-14 | 2020-05-29 | 上海恩睦信息科技有限公司 | Parkinson prediction method and device |
| CN111415746A (en) * | 2020-04-22 | 2020-07-14 | 上海邦邦机器人有限公司 | Physical function evaluation model generation method, physical function evaluation method, and physical function evaluation apparatus |
| CN111528842A (en) * | 2020-05-26 | 2020-08-14 | 复嶂环洲生物科技(上海)有限公司 | Quantitative assessment method for Parkinson disease symptoms based on physiological and behavioral indexes |
| CN111528842B (en) * | 2020-05-26 | 2023-01-03 | 复嶂环洲生物科技(上海)有限公司 | Quantitative assessment method for Parkinson disease symptoms based on physiological and behavioral indexes |
| CN112826504A (en) * | 2021-01-07 | 2021-05-25 | 中新国际联合研究院 | Games-based Parkinson symptom grade assessment method and device |
| CN112826504B (en) * | 2021-01-07 | 2024-03-26 | 中新国际联合研究院 | Game parkinsonism grade assessment method and device |
| CN117643456A (en) * | 2024-01-29 | 2024-03-05 | 北京航空航天大学 | Auxiliary evaluation system, method and storage medium for parkinsonism |
| CN117643456B (en) * | 2024-01-29 | 2024-05-24 | 北京航空航天大学 | Auxiliary evaluation system, method and storage medium for parkinsonism |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107157450B (en) | 2020-03-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107157450A (en) | Quantitative estimation method and system are carried out for the hand exercise ability to patient Parkinson | |
| Rigas et al. | Assessment of tremor activity in the Parkinson’s disease using a set of wearable sensors | |
| Arora et al. | High accuracy discrimination of Parkinson's disease participants from healthy controls using smartphones | |
| Printy et al. | Smartphone application for classification of motor impairment severity in Parkinson's disease | |
| CN104398263B (en) | A kind of disturbances in patients with Parkinson disease based on approximate entropy and mutual approximate entropy tremble symptom quantify evaluating method | |
| Burns et al. | Upper limb movement classification via electromyographic signals and an enhanced probabilistic network | |
| CN103054585B (en) | Biological motion information based upper limb shoulder elbow wrist joint motion function evaluation method | |
| Zhang et al. | Beyond the standard clinical rating scales: fine-grained assessment of post-stroke motor functionality using wearable inertial sensors | |
| CN108305680A (en) | Intelligent parkinsonism aided diagnosis method based on multi-element biologic feature and device | |
| Samuel et al. | Examining the effect of subjects' mobility on upper-limb motion identification based on EMG-pattern recognition | |
| CN111401435B (en) | Human body motion mode identification method based on motion bracelet | |
| McCool et al. | Lower arm electromyography (EMG) activity detection using local binary patterns | |
| CN114358194A (en) | Gesture tracking based detection method for abnormal limb behaviors of autism spectrum disorder | |
| Mitchell et al. | Automatically detecting asymmetric running using time and frequency domain features | |
| Pereira et al. | Using inertial sensors to evaluate exercise correctness in electromyography-based home rehabilitation systems | |
| Erim et al. | Decomposition of intramuscular EMG signals using a heuristic fuzzy expert system | |
| Raghu et al. | Automated biomedical signal quality assessment of electromyograms: Current challenges and future prospects | |
| Dyshel et al. | Quantifying levodopa-induced dyskinesia using depth camera | |
| Sejdic et al. | Selective regional correlation for pattern recognition | |
| Veer | A flexible approach for segregating physiological signals | |
| Eun et al. | Development of personalized urination recognition technology using smart bands | |
| CN116484290A (en) | Depression recognition model construction method based on Stacking integration | |
| CN111297366A (en) | Data processing method and diagnosis device for assisting disease diagnosis based on daily necessities | |
| Jain et al. | EMG Signal Feature Extraction, Normalization and Classification for Pain and Normal Muscles Using Genetic Algorithm and Support Vector Machine. | |
| CN110755084B (en) | Motion function assessment method and device based on active-passive and staged actions |
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 |