CN111990967A - Gait-based Parkinson disease recognition system - Google Patents
Gait-based Parkinson disease recognition system Download PDFInfo
- Publication number
- CN111990967A CN111990967A CN202010634516.7A CN202010634516A CN111990967A CN 111990967 A CN111990967 A CN 111990967A CN 202010634516 A CN202010634516 A CN 202010634516A CN 111990967 A CN111990967 A CN 111990967A
- Authority
- CN
- China
- Prior art keywords
- data
- gait
- module
- sensors
- identification system
- 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
- 208000018737 Parkinson disease Diseases 0.000 title claims abstract description 26
- 230000005021 gait Effects 0.000 title claims abstract description 19
- 238000004364 calculation method Methods 0.000 claims abstract description 21
- 238000013145 classification model Methods 0.000 claims abstract description 21
- 230000001133 acceleration Effects 0.000 claims abstract description 13
- 238000003745 diagnosis Methods 0.000 claims abstract description 6
- 238000010801 machine learning Methods 0.000 claims abstract description 6
- 238000013528 artificial neural network Methods 0.000 claims abstract description 4
- 230000000007 visual effect Effects 0.000 claims abstract description 4
- 238000000034 method Methods 0.000 claims description 17
- 238000001914 filtration Methods 0.000 claims description 12
- 238000000354 decomposition reaction Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 230000009467 reduction Effects 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 238000002790 cross-validation Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 11
- 206010044565 Tremor Diseases 0.000 description 5
- 210000002683 foot Anatomy 0.000 description 4
- 201000006517 essential tremor Diseases 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000000243 solution Substances 0.000 description 2
- 102000006441 Dopamine Plasma Membrane Transport Proteins Human genes 0.000 description 1
- 108010044266 Dopamine Plasma Membrane Transport Proteins Proteins 0.000 description 1
- 102000015554 Dopamine receptor Human genes 0.000 description 1
- 108050004812 Dopamine receptor Proteins 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 210000003423 ankle Anatomy 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- 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
- A61B5/103—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
- A61B5/112—Gait analysis
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Neurology (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- Neurosurgery (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Physiology (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Developmental Disabilities (AREA)
- Dentistry (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention relates to the technical field of medical diagnosis, and particularly discloses a gait-based Parkinson disease identification system, which comprises a sensor module, a calculation module and an interface module, wherein the sensor module, the calculation module and the interface module are electrically connected with each other; the sensor module comprises three triaxial acceleration sensors and three triaxial gyroscope sensors, and the three triaxial acceleration sensors and the three triaxial gyroscope sensors are correspondingly worn at the left hand, the right hand and the left foot of a user; the calculation module comprises a classification model, the classification model comprises a plurality of machine learning classification models and a deep neural network mLSTM classification model, and calculation is carried out according to sensor data; the interface module comprises a front end and a rear end and provides visual experience for a user; the invention can assist doctors to judge, and the doctors can more effectively judge whether patients suffer from Parkinson's disease.
Description
Technical Field
The invention relates to the technical field of medical diagnosis, in particular to a gait-based Parkinson disease identification system.
Background
Parkinson tremor (PD) and Essential Tremor (ET) are the two most common pathological tremors. However, because they have many similar features in clinical and physiological signal characteristics, the distinction between the two has been a difficult problem in clinical diagnosis. To avoid misdiagnosis, some hospitals use dopamine transporter imaging to determine whether parkinson tremor is present through dopamine receptor status. However, this method requires complex instrumentation and injection of radiotracers, which is time consuming and therefore can only be performed in a few hospital centers.
Therefore, the research of finding a simple and feasible method for auxiliary diagnosis of the Parkinson's disease and the Parkinson's disease is always a clinical research hotspot, and at present, many methods for identifying the Parkinson's disease based on machine learning have higher requirements on characteristics, so that once the Parkinson's disease and the Parkinson's disease are separated from the original data set, the accuracy rate is difficult to ensure. Therefore, the deep learning model based on the mLSTM is provided, and the gait signals of the Parkinson patients can be processed more effectively.
Disclosure of Invention
The present invention is directed to a gait-based parkinson's disease recognition system to solve the above-mentioned problems of the background art.
In order to achieve the purpose, the invention provides the following technical scheme: a gait-based Parkinson disease recognition system comprises a sensor module, a calculation module and an interface module, wherein the sensor module, the calculation module and the interface module are electrically connected with each other;
the sensor module comprises three triaxial acceleration sensors and three triaxial gyroscope sensors, and the three triaxial acceleration sensors and the three triaxial gyroscope sensors are correspondingly worn at the left hand, the right hand and the left foot of a user;
the calculation module comprises a classification model, the classification model comprises a plurality of machine learning classification models and a deep neural network mLSTM classification model, and calculation is carried out according to sensor data;
the interface module comprises a front end and a rear end and provides visual experience for a user.
Preferably, the invention also provides a use method of the gait-based parkinson's disease identification system, which comprises the following steps:
s1: acquiring user data;
s2: preprocessing the data;
s3: carrying out filtering processing on the data;
s4: expanding the data samples;
s5: carrying out feature extraction on the data;
s6: classifying the data result;
s7: and making a judgment according to the data.
Preferably, in the step S1, the user wears the sensors in the sensor modules at the left hand, the right hand and the left foot, and walks back and forth on a walkway with a length of 9-12 meters, wherein the number of times of walking back and forth is 3-5; the sensors in the sensor module acquire data and transmit the data to the computing module.
Preferably, the data acquired by the sensor module is eighteen-dimensional time sequence data, the acquired data is time-stamped and is data in millisecond level, and the acquisition frequency is 1000 HZ.
Preferably, in step S2, the duplicate data is deduplicated, the calculation module identifies three groups of sensor timestamps in the sensor module, selects the largest one of the start timestamps as the base time, deletes the data before the base time in the remaining two groups of sensors, and aligns the time axes.
Preferably, in step S3, the calculating module selects a DB4 wavelet basis as a mother wavelet to perform discrete wavelet transform on the original data, delete the high-frequency part in each layer, and continue to decompose the low-frequency part into high-frequency and low-frequency parts, each time reducing the frequency by one half; setting the Level to be 5, deleting high-frequency parts from the Level1 to the Level5, and only keeping the parts left after five times of wavelet decomposition.
Preferably, in step S4, the average value of all values is calculated first, the average value is subtracted from all points to make the waveform oscillate near the X axis, then the median filtering is performed on the samples to filter out the glitches, then all nodes are traversed, recording is started from the portion where the acceleration exceeds 0.5m/S2 until the portion less than 0.5m/S2 is ended, thereby obtaining the window, finally the samples with too short window are filtered out, and the extended samples are obtained in the window.
Preferably, in step S5, the difference between the maximum and minimum values of the data, the maximum value, the average absolute value, the root mean square, the standard deviation, the skewness, the number of zero points, and the coefficient of variation are selected as features; the sensor module is provided with six sensors, data are obtained in three directions of XYZ axes, and nine sensors are lifted for each group of dataObtaining 162-dimensional data in total; using a dimension reduction formula to reduce the dimension of the data, and finally obtaining 54 groups of data; the dimensionality reduction formula is as follows:
preferably, in the step S6, cross validation is applied to randomly select data from the sample as a test set, and the rest data are used as a training set, wherein the ratio of the test set to the training set is 1: 1-2; normalizing the characteristics to enable the expectation of the sample to be 0 and the variance to be 1, and classifying by using a classification model; the classification models comprise SVM, XGboost, Adaboost, LDA and LR models.
Preferably, in the step S7, based on the conclusion of the data in the step S6, the diagnosis content is given.
Compared with the prior art, the invention has the beneficial effects that: the system provided by the invention is characterized in that the left hand, the right hand and the left foot of a patient are provided with sensors, the motion data of the user during walking is detected, the data is processed by the computing module, then the corresponding model is selected, the model can give a judgment result, and a doctor can assist in diagnosing whether the patient suffers from the Parkinson's disease or not according to the output of the model; by applying the system, doctors can more effectively judge whether patients suffer from the Parkinson disease.
Drawings
FIG. 1 is a schematic structural view of the present invention;
FIG. 2 is a schematic diagram of data duplication processing according to an embodiment;
FIG. 3 is a graph of start time stamp data for three sensors in an embodiment;
FIG. 4 is a schematic diagram of an alignment time axis in the embodiment;
FIG. 5 is a diagram illustrating an exemplary discrete wavelet decomposition;
FIG. 6 is a diagram of five-layer discrete wavelet transform in an embodiment;
FIG. 7 is a schematic diagram showing comparison between before and after filtering in the embodiment;
FIG. 8 is a schematic diagram of the embodiment after decomposition of high and low frequencies;
FIG. 9 is a schematic diagram of an exemplary windowing sample;
FIG. 10 is a diagram illustrating the operation result of the SVM in the embodiment;
FIG. 11 is a diagram illustrating an operation result of an XGboost operation result in the embodiment;
FIG. 12 is a diagram illustrating the result of AdaBoost operation in an embodiment;
FIG. 13 shows LR run results in examples;
FIG. 14 shows the LDA operation result in the embodiment;
FIG. 15 is a diagram illustrating a use screen of the present invention;
reference numbers in the figures: 1. a sensor module; 2. a calculation module; 3. and an interface module.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, the present invention provides a technical solution: a gait-based Parkinson disease recognition system comprises a sensor module 1, a calculation module 2 and an interface module 3, wherein the sensor module 1, the calculation module 2 and the interface module 3 are electrically connected with each other;
the sensor module 1 comprises three triaxial acceleration sensors and three triaxial gyroscope sensors, and the three triaxial acceleration sensors and the three triaxial gyroscope sensors are correspondingly worn on the left hand, the right hand and the left foot of a user;
the calculation module 2 comprises a classification model, the classification model comprises a plurality of machine learning classification models and a deep neural network mLSTM classification model, and calculation is carried out according to sensor data;
the interface module 3 comprises a front end and a back end, and provides visual experience for users.
Example (b):
and (3) acquiring data: the method comprises the following steps that three-axis acceleration sensors and three-axis gyroscope sensors are worn on the left arm, the right arm and the left ankle of a patient, the patient walks back and forth by three to four times on a linear walkway with the length of about 10 meters, and time sequence data of eighteen dimensions in total are obtained; the acquired data is time-stamped, and is in millisecond level, and the acquisition frequency is 1000 HZ.
Meanwhile, the medical record of picture photographing can be used for obtaining the information of height, sex, age and the like of the patient. And additionally finding 10 normal persons as a comparison group, and collecting data in the same way.
Preprocessing of data: as shown in fig. 2, the acquired data includes many repeated data, all the data including the timeframe are communicated, and it is considered as dirty data, only one of the data is reserved, and the redundant data is deleted and subjected to deduplication processing.
As shown in fig. 3, since three sensors collectively provide data arranged in time series, the three sensors sometimes have a problem of time axis inconsistency, and the lack of data of individual sensors for a certain period of time may make calculation impossible. At this time, the largest one of the start timestamps of the three sensors is taken, which is 17866 in fig. 3; the data in the other two sets of data before 17866 are then deleted and the time axes are aligned as shown in FIG. 4.
And (3) filtering the data: wavelet Transform (WT) is commonly used in the field of signal processing, and this technique is very effective in dealing with problems in the time and frequency domains. The WT uses low and high pass filters to remove high and low frequency signal components and repeats the process of filtering the signal a set number of times by separating the signals and re-filtering the signal resulting from the previous filtering.
Signal decomposition using WT as shown in fig. 5, an input original signal passes through a high-pass filter to obtain a high-frequency signal component, and then passes through a down-sampling filter to be the high-frequency output of the Level. The low-frequency signals are the same, the next layer takes the low-frequency signal of the layer as input and decomposes the low-frequency signal into high-frequency and low-frequency signals again, and so on, and the whole structure is shown in figure 6.
And selecting a DB4 wavelet basis as a mother wavelet to perform discrete wavelet transform on the original signal, deleting the high-frequency part in each layer, and continuously decomposing the low-frequency part into the high-frequency part and the low-frequency part, wherein the frequency is reduced by one half each time. Setting the Level to be 5, deleting high-frequency parts from the Level1 to the Level5, and only keeping the parts left after five times of wavelet decomposition.
Taking one of the sensor accelerometers as an example, the results before and after wavelet transformation are shown in fig. 7, the high-frequency and low-frequency parts of the original signal after 5-layer discrete wavelet decomposition are shown in fig. 8, and it can be seen from the figure that the high-frequency part is mainly noise data with an average value of 0, so that the high-frequency part is simply removed, and the low-frequency part is reserved, so as to achieve the purpose of removing dryness.
By the filtering operation, the high frequency part of tremor can be removed, because it is known in the common sense that the part of human hand with tremor too high can be basically considered as noise data, and the subsequent processing can be adversely affected. Therefore, by filtering the noise, the accuracy is improved and the calculation amount is reduced.
Expansion of data samples: as shown in fig. 9, the samples are added in a windowed fashion. From the recorded live patient walking video and the generated acceleration waveform, it can be known that a "valley-peak-valley" of the patient is considered that the patient walks from one side to the other side, and the part with small amplitude in the middle is the noise data generated when the patient turns. In summary, the inverted portion can be deleted and each "valley-peak-valley" can be treated as a separate sample.
The average of all values is first calculated and subtracted from all points to oscillate the waveform around the X-axis. And then performing median filtering on the sample to filter out burrs. All nodes are traversed, recording starting from the part where the acceleration exceeds 0.5 m/(s) and ending up in the part less than 0.5 m/(s), thus obtaining the windowing. Finally, the samples with too short a window are filtered out, and a total of 190 samples are obtained through the window.
And (3) feature extraction of data:
selecting a series of standard characteristics: maximum-minimum difference, maximum, mean absolute, root mean square, standard deviation, skewness, number of zeros, and varianceAnd (5) obtaining 9-dimensional data by different coefficients. The sensor module 1 has six-dimensional attributes of three accelerometer sensors and three gyroscope sensors, and each sensor obtains data in the directions of three axes XYZ, so that 9 × 6 × 3 ═ 162 dimensional data is finally obtained; in order to reduce the dimensionality of data, a three-axis accelerometer and a three-axis gyroscope are considered to respectively generate three-phase data, the data of the three axes XYZ are the data of the same sensor, the wearing positions are the same, the data of the three axes XYZ can be considered to be the data with the same frequency and different amplitudes, and the square root of the data of the three axes X, Y and Z can be taken for dimensionality reduction; the dimension reduction processing formula is as follows:
classification of data: randomly extracting 76 groups of data as a test set from 190 groups of samples by using cross validation, and taking the rest 114 groups of data as a training set, wherein the ratio of the test set to the training set is 1: 1.5; normalizing the features so that the expectation of the sample is 0 and the variance is 1; a classification model in machine learning is used for comparison experiments, and an SVM, XGboost, Adaboost, LR and LDA five models are used for classifying results.
The operation results of the five models of the SVM, XGboost, Adaboost, LR and LDA are shown in fig. 10-14, and the comparison of the current models shows that the SVM has better overall accuracy and stability, and has the accuracy as high as 97 percent, and is enough for medical experiments.
The doctor makes a diagnosis based on the output data.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (10)
1. A gait-based parkinson's disease identification system, characterized by: the device comprises a sensor module (1), a calculation module (2) and an interface module (3), wherein the sensor module (1), the calculation module (2) and the interface module (3) are electrically connected with each other;
the sensor module (1) comprises three triaxial acceleration sensors and three triaxial gyroscope sensors, and the three triaxial acceleration sensors and the three triaxial gyroscope sensors are correspondingly worn on the left hand, the right hand and the left foot of a user;
the calculation module (2) comprises a classification model, the classification model comprises a plurality of machine learning classification models and a deep neural network mLSTM classification model, and calculation is carried out according to sensor data;
the interface module (3) comprises a front end and a rear end, and visual experience is provided for a user.
2. The method of using a gait-based parkinson's disease identification system according to claim 1, characterized in that: the method comprises the following steps:
s1: acquiring user data;
s2: preprocessing the data;
s3: carrying out filtering processing on the data;
s4: expanding the data samples;
s5: carrying out feature extraction on the data;
s6: classifying the data result;
s7: and making a judgment according to the data.
3. The method of using a gait-based parkinson's disease identification system according to claim 2, characterized in that: in the step S1, the left hand, the right hand and the left foot of a user wear the sensors in the sensor module (1), and the user walks back and forth on a walking path with the length of 9-12 meters, wherein the walking times are 3-5 times; the sensors in the sensor module (1) acquire data and transmit the data to the computing module (2).
4. The method of using a gait-based parkinson's disease identification system according to claim 3, characterized in that: the data acquired by the sensor module (1) are eighteen-dimensional time sequence data, the acquired data are provided with time stamps and are data in millisecond level, and the acquisition frequency is 1000 HZ.
5. The method of using a gait-based parkinson's disease identification system according to claim 2, characterized in that: in step S2, the duplicate data is deduplicated, the calculation module (2) identifies three sets of sensor timestamps in the sensor module (1), selects the largest one of the start timestamps as the base time, deletes the data before the base time in the remaining two sets of sensors, and aligns the time axis.
6. The method of using a gait-based parkinson's disease identification system according to claim 2, characterized in that: in the step S3, the calculating module (2) selects a DB4 wavelet basis as a mother wavelet to perform discrete wavelet transform on the original data, deletes a high-frequency part in each layer, and continuously decomposes a low-frequency part into a high-frequency part and a low-frequency part, each time reducing the frequency by one half; setting the Level to be 5, deleting high-frequency parts from the Level1 to the Level5, and only keeping the parts left after five times of wavelet decomposition.
7. The method of using a gait-based parkinson's disease identification system according to claim 2, characterized in that: in step S4, the average of all the values is calculated first, the average is subtracted from all the points to make the waveform oscillate near the X-axis, then the samples are median filtered to filter out the glitches, and then all the nodes are traversed, and the acceleration exceeds 0.5m/S2Until less than 0.5m/s2And (4) finishing the part to obtain a window, finally filtering out samples with too short window, and obtaining expanded samples in the window.
8. The method of using a gait-based parkinson's disease identification system according to claim 2, characterized in that: in step S5, selecting the difference between the maximum and minimum values, the maximum value, the average absolute value, the root mean square, the standard deviation, the skewness, the number of zero points, and the coefficient of variation of the data as features; the sensingThe device module (1) is provided with six sensors, data are obtained in three directions of XYZ axes, nine features are lifted in each group of data, and 162-dimensional data are obtained in total; using a dimension reduction formula to reduce the dimension of the data, and finally obtaining 54 groups of data; the dimensionality reduction formula is as follows:
9. the method of using a gait-based parkinson's disease identification system according to claim 2, characterized in that: in the step S6, randomly selecting data from a sample as a test set by applying cross validation, and using the rest data as a training set, wherein the ratio of the test set to the training set is 1: 1-2; normalizing the characteristics to enable the expectation of the sample to be 0 and the variance to be 1, and classifying by using a classification model; the classification models comprise SVM, XGboost, Adaboost, LDA and LR models.
10. The method of using a gait-based parkinson's disease identification system according to claim 2, characterized in that: in the step S7, based on the conclusion of the data in the step S6, the diagnosis content is given.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010634516.7A CN111990967A (en) | 2020-07-02 | 2020-07-02 | Gait-based Parkinson disease recognition system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010634516.7A CN111990967A (en) | 2020-07-02 | 2020-07-02 | Gait-based Parkinson disease recognition system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111990967A true CN111990967A (en) | 2020-11-27 |
Family
ID=73466422
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010634516.7A Pending CN111990967A (en) | 2020-07-02 | 2020-07-02 | Gait-based Parkinson disease recognition system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111990967A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112674762A (en) * | 2020-12-28 | 2021-04-20 | 江苏省省级机关医院 | Parkinson tremble evaluation device based on wearable inertial sensor |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106971059A (en) * | 2017-03-01 | 2017-07-21 | 福州云开智能科技有限公司 | A kind of wearable device based on the adaptive health monitoring of neutral net |
| CN108514421A (en) * | 2018-03-30 | 2018-09-11 | 福建幸福家园投资管理有限公司 | The method for promoting mixed reality and routine health monitoring |
| CN109069066A (en) * | 2015-09-28 | 2018-12-21 | 卡斯西部储备大学 | Wearable and connection gait analysis system |
| WO2019036805A1 (en) * | 2017-08-22 | 2019-02-28 | Orpyx Medical Technologies Inc. | Method and system for activity classification |
| CN109805935A (en) * | 2017-11-21 | 2019-05-28 | 北京周智物联科技有限公司 | A kind of intelligent waistband based on artificial intelligence hierarchical layered motion recognition method |
| US20190172232A1 (en) * | 2017-12-05 | 2019-06-06 | Samsung Eletrônica da Amazônia Ltda. | Method and system for sensor data recognition using data enrichment for the learning process |
| CN110522456A (en) * | 2019-09-26 | 2019-12-03 | 安徽中医药大学 | A kind of WD based on deep learning trembles conditions of patients self-evaluating system |
| CN110638458A (en) * | 2019-08-26 | 2020-01-03 | 广东省人民医院(广东省医学科学院) | Gait data-based rehabilitation training effect evaluation method and device |
| CN111225612A (en) * | 2017-10-17 | 2020-06-02 | 萨蒂什·拉奥 | Neural obstacle identification and monitoring system based on machine learning |
-
2020
- 2020-07-02 CN CN202010634516.7A patent/CN111990967A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109069066A (en) * | 2015-09-28 | 2018-12-21 | 卡斯西部储备大学 | Wearable and connection gait analysis system |
| CN106971059A (en) * | 2017-03-01 | 2017-07-21 | 福州云开智能科技有限公司 | A kind of wearable device based on the adaptive health monitoring of neutral net |
| WO2019036805A1 (en) * | 2017-08-22 | 2019-02-28 | Orpyx Medical Technologies Inc. | Method and system for activity classification |
| CN111225612A (en) * | 2017-10-17 | 2020-06-02 | 萨蒂什·拉奥 | Neural obstacle identification and monitoring system based on machine learning |
| CN109805935A (en) * | 2017-11-21 | 2019-05-28 | 北京周智物联科技有限公司 | A kind of intelligent waistband based on artificial intelligence hierarchical layered motion recognition method |
| US20190172232A1 (en) * | 2017-12-05 | 2019-06-06 | Samsung Eletrônica da Amazônia Ltda. | Method and system for sensor data recognition using data enrichment for the learning process |
| CN108514421A (en) * | 2018-03-30 | 2018-09-11 | 福建幸福家园投资管理有限公司 | The method for promoting mixed reality and routine health monitoring |
| CN110638458A (en) * | 2019-08-26 | 2020-01-03 | 广东省人民医院(广东省医学科学院) | Gait data-based rehabilitation training effect evaluation method and device |
| CN110522456A (en) * | 2019-09-26 | 2019-12-03 | 安徽中医药大学 | A kind of WD based on deep learning trembles conditions of patients self-evaluating system |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112674762A (en) * | 2020-12-28 | 2021-04-20 | 江苏省省级机关医院 | Parkinson tremble evaluation device based on wearable inertial sensor |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112353407B (en) | Evaluation system and method based on active training of neurological rehabilitation | |
| CN114010171B (en) | Classifier setting method based on heartbeat data | |
| Kee et al. | Binary and multi-class motor imagery using Renyi entropy for feature extraction | |
| Zhao et al. | An IoT-based wearable system using accelerometers and machine learning for fetal movement monitoring | |
| CN111839506B (en) | Mental load detection method and device | |
| CN109044280B (en) | Sleep staging method and related equipment | |
| Naseer et al. | Classification of normal and abnormal ECG signals based on their PQRST intervals | |
| Talatov et al. | Algorithmic and software analysis and processing of ECG signals | |
| Gasparini et al. | A deep learning approach to recognize cognitive load using ppg signals | |
| CN111990967A (en) | Gait-based Parkinson disease recognition system | |
| CN113609975A (en) | Modeling method for tremor detection, hand tremor detection device and method | |
| CN116784860B (en) | Electrocardiosignal characteristic extraction system based on morphological heart beat template clustering | |
| Özel et al. | Implementation of Artifact Removal Algorithms in Gait Signals for Diagnosis of Parkinson Disease. | |
| Dubey et al. | A neural network approach for ECG classification | |
| Hasan et al. | Preliminary study on real-time prediction of gait acceleration intention from volition-associated EEG patterns | |
| KR102598219B1 (en) | Method and apparatus for restoring remote photoplethysmography | |
| CN115736920A (en) | Depression state identification method and system based on bimodal fusion | |
| Narayan | Analysis of MLP and DSLVQ classifiers for EEG signals based movements identification | |
| CN108537200B (en) | Apparatus and method for selectively collecting electroencephalographic data through motion recognition | |
| KR20220158462A (en) | EMG signal-based recognition information extraction system and EMG signal-based recognition information extraction method using the same | |
| Manjula et al. | BCG Artifact Removal Using Improved Independent Component Analysis Approach | |
| Aravind et al. | ECG Classification and Arrhythmia Detection Using Wavelet Transform and Convolutional Neural Network | |
| Dong et al. | Home-based detection of epileptic seizures using a bracelet with motor sensors | |
| Bhagwat et al. | Human disposition detection using EEG signals | |
| Yosrita et al. | Denoising of eeg signal based on word imagination using ica for artifact and noise removal on unspoken speech |
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: 20201127 |
|
| RJ01 | Rejection of invention patent application after publication |