CN112075940A - Tremor detection system based on bidirectional long-time and short-time memory neural network - Google Patents
Tremor detection system based on bidirectional long-time and short-time memory neural network Download PDFInfo
- Publication number
- CN112075940A CN112075940A CN202010996323.6A CN202010996323A CN112075940A CN 112075940 A CN112075940 A CN 112075940A CN 202010996323 A CN202010996323 A CN 202010996323A CN 112075940 A CN112075940 A CN 112075940A
- Authority
- CN
- China
- Prior art keywords
- data
- tremor
- hand
- triaxial acceleration
- neural network
- 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
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—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
- A61B5/1101—Detecting tremor
-
- 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
- A61B5/1121—Determining geometric values, e.g. centre of rotation or angular range of movement
-
- 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
-
- 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
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Abstract
The utility model provides a tremble detecting system based on two-way long-and-short term memory neural network, relates to the artificial intelligence field, detects to tremble among the prior art and has the problem that the accuracy is low, include: the hand detection module and the model prediction module; the hand detection module comprises a tremor data acquisition unit and a tremor data processing unit; the tremor data acquisition unit is used for acquiring hand triaxial acceleration signals; the tremor data processing unit is used for converting the hand triaxial acceleration signal into hand triaxial acceleration data; the model prediction module comprises a hand tremor data processing unit and a model training unit; the hand tremor data processing unit is used for processing the received triaxial acceleration data to obtain a training set and a test set; and the model training unit is used for training the model by utilizing the training set and the test set to obtain the trained model. The invention is used for tremor detection and has high detection efficiency.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a tremor detection system based on a bidirectional long-time and short-time memory neural network.
Background
Hand tremor is a common clinical manifestation, and is seen in various nervous system diseases, including essential tremor, parkinson's disease, psychogenic tremor, dystonic tremor, cerebellar tremor, metabolic tremor, peripheral neuropathic tremor, hepatolenticular degeneration and other diseases. Although the tremor frequency and amplitude in different diseases are different, for example, essential tremor is usually low in amplitude and fast in frequency (8-10Hz), and the tremor frequency of Parkinson's disease is usually 4-6Hz, the current tremor detection methods are easily affected by subjective factors of detectors, and the assessment cannot be objective and accurate, the sensitivity is not high, and the tiny change cannot be detected.
The development of machine learning technology in recent years provides an important means for analyzing hand tremor. Methods for detecting resting tremor have been previously proposed, but there is a lack of relevant technical means for detecting tremor during specific movements. The tremor signal of the tested person in daily life can be monitored to provide early warning for the tested person and prevent the tested person. In addition, the judgment process can be simplified by predicting through machine learning, the detection accuracy is improved, and the method has practical significance.
Disclosure of Invention
The purpose of the invention is: aiming at the problem of low accuracy in tremor detection in the prior art, a tremor detection system based on a bidirectional long-time and short-time memory neural network is provided.
The technical scheme adopted by the invention to solve the technical problems is as follows:
a tremor detection system based on a bi-directional long-and-short-term memory neural network, comprising: the hand detection module and the model prediction module;
the hand detection module comprises a tremor data acquisition unit and a tremor data conversion unit;
the tremor data acquisition unit is used for acquiring hand triaxial acceleration signals;
the tremor data conversion unit is used for converting the hand triaxial acceleration signal into hand triaxial acceleration data;
the model prediction module comprises a hand tremor data processing unit and a model training unit;
the hand tremor data processing unit is used for processing the received triaxial acceleration data to obtain a training set and a test set;
the hand tremor data processing unit comprises a hand tremor data processing unit and a model training unit, wherein the model training unit is used for training a model by utilizing a training set and a testing set to obtain the trained model, and the hand tremor data processing unit comprises the following specific steps:
the method comprises the following steps: performing qualification detection on the received triaxial acceleration data, eliminating unqualified triaxial acceleration data, and keeping qualified triaxial acceleration data;
step two: cutting qualified triaxial acceleration data into uniform size;
step three: performing data fusion on the cut triaxial acceleration data;
step four: the data after data fusion is given to a label and is divided into two types of data with tremor and data without tremor;
step five: splicing the two types of data into a complete data set;
step six: randomly disordering the sequence of the spliced data;
step seven: and carrying out normalization processing on the disordered data.
Further, the qualification testing specifically comprises the following steps:
the method comprises the following steps: setting a threshold range j;
the first step is: collecting three groups of hand triaxial acceleration data and obtaining the fusion acceleration of the three groups of data;
step one is three: judging whether the three groups of data fusion accelerations are within a threshold range j or not; if the current value is within the threshold value range j, the qualification inspection is judged to be passed, and if the current value is not within the threshold value range j, the qualification inspection is judged not to be passed.
Further, the value range of the threshold value range j is as follows:
10m/s2＜j≤20m/s2。
further, the threshold value range j is 15m/s2。
Further, the tremor data acquisition unit acquires hand triaxial acceleration signals by using a gyroscope sensor.
Further, the data fusion specifically comprises the following steps:
recording the three-axis acceleration of a when the gyroscope sensor is staticx0、ay0、az0Fusing the triaxial acceleration in the data passing the qualification detection, and setting the triaxial acceleration in the data set as ax、ay、azThen the fusion formula is:
where a is the post-fusion acceleration.
Furthermore, the model is built based on the bidirectional LSTM and comprises three one-dimensional convolutional layers, two BiLSTM layers, a full connection layer and an output layer, the number of neurons in the two BiLSTM layers is 128, dropout is added into the full connection layer, the output layer is softmax, and RMSprop is adopted by the optimizer.
Further, the activation function of the model is a ReLU function, and the ReLU function is expressed as:
where x is the input to the previous layer.
Further, the system also comprises a control unit, and the control unit is used for controlling the turning-on and turning-off of the tremor data acquisition unit.
Further, the system further comprises a visualization unit, wherein the visualization unit is used for displaying the hand triaxial acceleration signal and the hand triaxial acceleration data.
The invention has the beneficial effects that:
the system can detect whether the tremor exists in the moving testee, has the advantages that compared with other classification models such as SVM and Logistic, the accuracy of the model on the test set is 0.851, and the accuracy of the Logistic model and the SVM model on the test set is 0.605 and 0.7 respectively, and is excellent in performance, high in detection accuracy and high in sensitivity from the performance of the test set.
In addition, for the same type of neural network, the difference in the structure of the neural network also affects the degree of excellence of the model. As shown in fig. 5, a network structure comprising three one-dimensional convolutional layers and two BiLSTM layers is used, which performs best.
Drawings
FIG. 1 is a tremor data processing flow diagram;
FIG. 2 is a tremor data identification network;
FIG. 3 is a flow chart of complete hand tremor diagnosis;
FIG. 4 is a graph comparing accuracy on test sets using different methods;
FIG. 5 is a graph comparing the accuracy of different neural networks on a test set.
Detailed Description
The first embodiment is as follows: specifically describing the present embodiment with reference to fig. 1, fig. 2 and fig. 3, the tremor detection system based on the bidirectional long-and-short term memory neural network according to the present embodiment includes: the hand detection module, the upper computer software module and the model prediction module.
The hand tremor detection of the invention is a fingertip tremor detection, and therefore the invention is carried out according to the following steps:
a. performing hand tremor detection activities by hand detection modules
The hand tremor detection module comprises a tremor data acquisition unit, a tremor data processing unit and an upper computer communication unit (serial port). Wherein, tremble data acquisition unit selects MPU6050 as the sensor, and tremble data processing unit selects STM 32F 103 as microcontroller, and the serial ports standard is RS-232.
When the tremor data acquisition unit acquires hand tremor data of a tested person, in order to enrich the data set, two preset motion states are detected, so that the tested person can execute the following two motion states:
firstly, the measured hand is kept still, and the other hand can swing reasonably;
and secondly, keeping the measured hand still and walking on a specified route.
The specific steps of detecting hand tremor aiming at the two states are as follows: the measured person wears a finger sleeve on the index finger of a left hand or a right hand, the main body part of the device is tied on the wrist, the body and the measured hand are kept still according to a motion state I, the other hand swings the arm back and forth slowly on the horizontal position, in order to reduce interference factors, a tester monitors the back and forth reciprocating speed of the measured person, controls the speed within a reasonable range, eliminates interference data of too fast motion and too slow motion, and records the 10-second triaxial acceleration data of the measured person in the motion process; then according to the second motion state, keeping the hand of the tested person still, designating a walking route by the tested person, monitoring the walking speed, controlling the walking speed within a reasonable range, and recording the 10-second triaxial acceleration data in the process; after the collection, the tested person takes a short rest and carries out the same detection twice again. The collected data are identified, processed and processed by the tremor data processing module and transmitted to the upper computer software module through the upper computer communication unit (serial port).
b. The upper computer software module saves and visualizes the information and data of the testee
The upper computer software module comprises a communication mode setting part, a test starting and ending control part, a three-axis acceleration display part, a data visualization part and a tested person information display part.
The hand tremor detection module communicates detected data with an upper computer software module through a serial port, firstly, a communication mode setting part of the upper computer software module sets a serial port, a baud rate and the like for communicating with a lower computer, and after communication is established, information of a detected person is filled in according to the prompt of the upper computer. The start button is clicked at the test start and end control part, at the moment, triaxial acceleration data generated by the movement of the tested person is displayed on the triaxial acceleration display part in real time, the waveform is stable, when no abnormity occurs, the start recording button is clicked, the upper computer automatically records 10-second data and stores the data in a txt file by taking the information of the tested person as a name, and the data can be conveniently processed in batches subsequently. After the data recording is finished, clicking a frequency spectrum button, automatically calling acceleration data of the tested person by the upper computer, and respectively visualizing the original data and the processed data on an original data oscillogram and a filtering oscillogram. Finally, the end button is clicked, and the relevant information of the tested person can be automatically updated on the information display part of the tested person.
c. Eligibility detection of hand tremor data
The hand tremor data acquisition module acquires data generated by the hand tremor detection activity of the tested person each time in real time, and then performs qualification inspection on the group of tremor data, and if the group of tremor data passes the qualification inspection, the group of tremor data is stored. Otherwise, the tremor data generated by the hand tremor detection activity is invalid, and the tested person executes the hand tremor detection activity again.
The qualification inspection is to judge whether the sum of squares of acceleration information included in the time series of three groups of tremor data obtained by three hand tremor detection activities of the tested person in the same motion state is in a preset range, and if so, the group of tremor data is judged to pass the qualification inspection, and the data is stored. Otherwise, judging that the set of tremor data fails the qualification check, and deleting the data, wherein the sum of the squares of the accelerations is A2The preset range is 100 < A2≤400。
The testees passing through the qualification detection are divided into testees with tremor and testees without tremor, and the data of the two testees passing through the qualification detection are respectively stored by the upper computer.
The present experiment tested 9 subjects who had tremor themselves and 9 subjects who did not have tremor, and the 18 participants had no difference in the number of sexes, and then the 18 participants were tested as described above.
d. Neural network model training based on stored data
The model predictive model comprises a hand tremor data processing part and a model training part.
The tremor data processing part comprises the following steps of firstly cutting the length of a data sequence into the same size, then carrying out data fusion on the triaxial acceleration of the data, then giving labels to two types of data stored by an upper computer, marking the data of a testee without tremor as 0, marking the data of the testee with tremor as 1, splicing the two types of data into a complete data set, randomly disordering the sequence of the spliced data, and finally carrying out normalization processing on the data and dividing a training set and a testing set.
The data fusion specifically comprises the following steps: recording the three-axis acceleration of the sensor at rest as ax0、ay0、az0. Fusing the three-axis acceleration in the screened data, and setting the three-axis acceleration in the data set as ax、ay、azThen the fusion formula is:
where a is the post-fusion acceleration.
The data normalization and the data division are performed, and the data normalization process is performed to eliminate dimension influence among indexes so as to solve comparability among data indexes. After the raw data are subjected to data standardization processing, all indexes are in the same order of magnitude, and the method is suitable for comprehensive comparison and evaluation. Data partitioning divides a data set into a training set and a testing set, and provides raw materials for model training.
The model training part is constructed based on the bidirectional LSTM and is a neural network formed by three one-dimensional convolution layers, two BiLSTM layers, a full connection layer and an output layer. The convolution kernel size, convolution step size, convolution kernel number, etc. of three one-dimensional convolution layers need to be set, the activation function is set to ReLU (corrected Linear Unit), and the mathematical formula of ReLU is:
wherein x is the input of the previous layer; the number of neurons in two BilSTM layers is set to be 128, and dropout is added in a full connection layer, so that the problem that overfitting is caused by excessively depending on a certain neuron in information transmission to influence generalization capability is solved. And finally, the output layer outputs the prediction result by softmax. The optimizer adopts RMSprop, wherein the learning rate alpha is 0.001, the attenuation rate beta is 0.9, and the fuzzy factor is 1 multiplied by 10-8The learning rate drop value of each iteration is 1 multiplied by 10-8。
Based on the data set, taking out 40% of processed data as a test set, taking the rest 60% as a training set, training the training set in the built model, storing the model with the minimum cost function (or less than a threshold) through repeated iteration, and finally detecting a training result by using the test set.
The system can be used for detecting hand tremor in practical application, can also be used for Parkinson patients and essential tremor patients in specific implementation, and can obtain a corresponding prediction result by inputting fingertip data of the patient into a model when other patients are detected.
It should be noted that the detailed description is only for explaining and explaining the technical solution of the present invention, and the scope of protection of the claims is not limited thereby. It is intended that all such modifications and variations be included within the scope of the invention as defined in the following claims and the description.
Claims (10)
1. A tremor detection system based on a bi-directional long-and-short-term memory neural network, comprising: the hand detection module and the model prediction module;
the hand detection module comprises a tremor data acquisition unit and a tremor data conversion unit;
the tremor data acquisition unit is used for acquiring hand triaxial acceleration signals;
the tremor data conversion unit is used for converting the hand triaxial acceleration signal into hand triaxial acceleration data;
the model prediction module comprises a hand tremor data processing unit and a model training unit;
the hand tremor data processing unit is used for processing the received triaxial acceleration data to obtain a training set and a test set;
the model training unit is used for training the model by utilizing a training set and a testing set to obtain a trained model, and is characterized in that the hand tremor data processing unit comprises the following specific steps:
the method comprises the following steps: performing qualification detection on the received triaxial acceleration data, eliminating unqualified triaxial acceleration data, and keeping qualified triaxial acceleration data;
step two: cutting qualified triaxial acceleration data into uniform size;
step three: performing data fusion on the cut triaxial acceleration data;
step four: the data after data fusion is given to a label and is divided into two types of data with tremor and data without tremor;
step five: splicing the two types of data into a complete data set;
step six: randomly disordering the sequence of the spliced data;
step seven: and carrying out normalization processing on the disordered data.
2. The tremor detection system based on the bidirectional long-and-short-term memory neural network according to claim 1, wherein the qualification detection specifically comprises the following steps:
the method comprises the following steps: setting a threshold range j;
the first step is: collecting three groups of hand triaxial acceleration data and obtaining the fusion acceleration of the three groups of data;
step one is three: judging whether the three groups of data fusion accelerations are within a threshold range j or not; if the current value is within the threshold value range j, the qualification inspection is judged to be passed, and if the current value is not within the threshold value range j, the qualification inspection is judged not to be passed.
3. The tremor detection system based on the bidirectional long-and-short-term memory neural network according to claim 2, wherein the threshold value range j has a value range of:
10m/s2＜j≤20m/s2。
4. the tremor detection system based on the bi-directional long-and-short-term memory neural network of claim 3, wherein the threshold range j is 15m/s2。
5. The tremor detection system based on a bidirectional long-and-short-term memory neural network of claim 4, wherein said tremor data acquisition unit utilizes a gyroscope sensor to acquire hand triaxial acceleration signals.
6. The tremor detection system based on the bidirectional long-and-short-term memory neural network according to claim 5, wherein the data fusion specifically comprises the following steps:
recording the three-axis acceleration of a when the gyroscope sensor is staticx0、ay0、az0Fusing the triaxial acceleration in the data passing the qualification detection, and setting the triaxial acceleration in the data set as ax、ay、azThen the fusion formula is:
where a is the post-fusion acceleration.
7. The tremor detection system based on a bidirectional long-and-short-term memory neural network according to claim 6, wherein the model is built based on bidirectional LSTM, and includes three one-dimensional convolution layers, two BiLSTM layers, a full connection layer and an output layer, the number of neurons in the two BiLSTM layers is 128, the full connection layer is added with dropout, the output layer is softmax, and the optimizer adopts RMSprop.
8. The tremor detection system based on a bi-directional long-and-short-term memory neural network of claim 7, wherein the activation function of the model is a ReLU function, and the ReLU function is expressed as:
where x is the input to the previous layer.
9. The tremor detection system based on a bidirectional long-and-short-term memory neural network according to claim 1, wherein the system further comprises a control unit for controlling the tremor data acquisition unit to be turned on and off.
10. The tremor detection system based on a bi-directional long-short term memory neural network of claim 1, wherein said system further comprises a visualization unit for displaying the hand triaxial acceleration signal and the hand triaxial acceleration data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010996323.6A CN112075940A (en) | 2020-09-21 | 2020-09-21 | Tremor detection system based on bidirectional long-time and short-time memory neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010996323.6A CN112075940A (en) | 2020-09-21 | 2020-09-21 | Tremor detection system based on bidirectional long-time and short-time memory neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112075940A true CN112075940A (en) | 2020-12-15 |
Family
ID=73739326
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010996323.6A Pending CN112075940A (en) | 2020-09-21 | 2020-09-21 | Tremor detection system based on bidirectional long-time and short-time memory neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112075940A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112926553A (en) * | 2021-04-25 | 2021-06-08 | 北京芯盾时代科技有限公司 | Training method and device for motion detection network |
| CN113017613A (en) * | 2021-03-03 | 2021-06-25 | 四川大学华西医院 | Artificial intelligence-based cardiac shock wave signal processing method and computer equipment |
| CN113705649A (en) * | 2021-08-20 | 2021-11-26 | 哈尔滨医科大学 | Hand tremor detection method and system based on EMD-SVD feature extraction |
| CN113822170A (en) * | 2021-08-31 | 2021-12-21 | 西安理工大学 | Hand tremor recognition method in non-static state of body |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102610056A (en) * | 2012-03-16 | 2012-07-25 | 清华大学 | Mobile phone wearing mode-oriented falling event detection system and method |
| CN102667672A (en) * | 2009-07-07 | 2012-09-12 | 韩铮 | Acceleration motion identify method and system thereof |
| CN103345627A (en) * | 2013-07-23 | 2013-10-09 | 清华大学 | Action recognition method and device |
| CN104731307A (en) * | 2013-12-20 | 2015-06-24 | 孙伯元 | Somatic action identifying method and man-machine interaction device |
| CN107493371A (en) * | 2016-06-13 | 2017-12-19 | 中兴通讯股份有限公司 | A kind of recognition methods, device and the terminal of the motion feature of terminal |
| CN109276255A (en) * | 2018-11-27 | 2019-01-29 | 平安科技（深圳）有限公司 | A kind of limb tremor detection method and device |
| CN110286774A (en) * | 2019-07-03 | 2019-09-27 | 中国科学技术大学 | A kind of sign Language Recognition Method based on Wrist-sport sensor |
| CN111544005A (en) * | 2020-05-15 | 2020-08-18 | 中国科学院自动化研究所 | Parkinson's disease dyskinesia quantification and identification method based on support vector machine |
| CN111680661A (en) * | 2020-06-19 | 2020-09-18 | 哈尔滨工业大学 | Mechanical rotating part performance degradation tracking method based on multi-feature fusion |
-
2020
- 2020-09-21 CN CN202010996323.6A patent/CN112075940A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102667672A (en) * | 2009-07-07 | 2012-09-12 | 韩铮 | Acceleration motion identify method and system thereof |
| CN102610056A (en) * | 2012-03-16 | 2012-07-25 | 清华大学 | Mobile phone wearing mode-oriented falling event detection system and method |
| CN103345627A (en) * | 2013-07-23 | 2013-10-09 | 清华大学 | Action recognition method and device |
| CN104731307A (en) * | 2013-12-20 | 2015-06-24 | 孙伯元 | Somatic action identifying method and man-machine interaction device |
| CN107493371A (en) * | 2016-06-13 | 2017-12-19 | 中兴通讯股份有限公司 | A kind of recognition methods, device and the terminal of the motion feature of terminal |
| CN109276255A (en) * | 2018-11-27 | 2019-01-29 | 平安科技（深圳）有限公司 | A kind of limb tremor detection method and device |
| CN110286774A (en) * | 2019-07-03 | 2019-09-27 | 中国科学技术大学 | A kind of sign Language Recognition Method based on Wrist-sport sensor |
| CN111544005A (en) * | 2020-05-15 | 2020-08-18 | 中国科学院自动化研究所 | Parkinson's disease dyskinesia quantification and identification method based on support vector machine |
| CN111680661A (en) * | 2020-06-19 | 2020-09-18 | 哈尔滨工业大学 | Mechanical rotating part performance degradation tracking method based on multi-feature fusion |
Non-Patent Citations (7)
| Title |
|---|
| 张雨倩等: "帕金森震颤与原发性震颤的计算机辅助诊断方法综述", 《计算机科学》 * |
| 张黎明: "智能运动控制训练结合综合康复对脑卒中患者平衡及下肢功能的影响", 《中国优秀硕士学位论文全文数据库 (医药卫生科技辑)》 * |
| 成娟: "基于表面肌电和加速度信号融合的动作识别和人体行为分析研究", 《中国博士学位论文全文数据库 (信息科技辑)》 * |
| 杨开天: "RAH-66环境振动合格性要求", 《直升机技术》 * |
| 梁雅庆: "帕金森患者震颤信号提取及量化方法研究", 《中国优秀硕士学位论文全文数据库 (医药卫生科技辑)》 * |
| 王洋等: "仿人机器人步态平衡泛化模型的建立与仿真", 《智能系统学报》 * |
| 鲁鹏威等: "基于MEMS的帕金森病人体姿态分类监测系统", 《现代电子技术》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113017613A (en) * | 2021-03-03 | 2021-06-25 | 四川大学华西医院 | Artificial intelligence-based cardiac shock wave signal processing method and computer equipment |
| CN112926553A (en) * | 2021-04-25 | 2021-06-08 | 北京芯盾时代科技有限公司 | Training method and device for motion detection network |
| CN113705649A (en) * | 2021-08-20 | 2021-11-26 | 哈尔滨医科大学 | Hand tremor detection method and system based on EMD-SVD feature extraction |
| CN113705649B (en) * | 2021-08-20 | 2024-01-12 | 哈尔滨医科大学 | EMD-SVD feature extraction-based hand tremor detection method and system |
| CN113822170A (en) * | 2021-08-31 | 2021-12-21 | 西安理工大学 | Hand tremor recognition method in non-static state of body |
| CN113822170B (en) * | 2021-08-31 | 2024-02-06 | 西安理工大学 | Method for identifying tremor of hands of body in non-stationary state |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112075940A (en) | Tremor detection system based on bidirectional long-time and short-time memory neural network | |
| RU2757048C1 (en) | Method and system for assessing the health of the human body based on the large-volume sleep data | |
| KR20200005987A (en) | System and method for diagnosing cognitive impairment using touch input | |
| CN106725376B (en) | Sign detection method and device | |
| CN110659677A (en) | Human body falling detection method based on movable sensor combination equipment | |
| US20210307686A1 (en) | Methods and systems to detect eating | |
| CN115249541A (en) | Dynamic blood glucose prediction model construction method based on deep learning | |
| CN107019501B (en) | Remote tumble detection method and system based on genetic algorithm and probabilistic neural network | |
| CN111991001A (en) | Motion data analysis method, system, equipment and readable storage medium | |
| Ali et al. | Human activity recognition system using smart phone based accelerometer and machine learning | |
| CN108962379B (en) | Mobile phone auxiliary detection system for cranial nerve system diseases | |
| CN112674760A (en) | Wearable sensor-based Parkinson upper limb movement detection method | |
| CN116739037A (en) | Personality model construction method and device with personality characteristics | |
| CN115336979A (en) | Multitask tremor automatic detection method based on wearable device and detection device thereof | |
| CN115444422A (en) | Eye movement data-based real environment psychological load assessment method and system | |
| Hallgrímsson et al. | Learning individualized cardiovascular responses from large-scale wearable sensors data | |
| CN115159286A (en) | Take advantage of terraced comfort level evaluation system | |
| CN114869272A (en) | Posture tremor detection model, posture tremor detection algorithm, and posture tremor detection apparatus | |
| Raychaudhuri et al. | Prescriptive analytics for impulsive behaviour prevention using real-time biometrics | |
| Yang et al. | Loneliness Forecasting Using Multi-modal Wearable and Mobile Sensing in Everyday Settings | |
| CN116019429A (en) | Health monitoring method, device, equipment and storage medium based on physiological index | |
| US10079074B1 (en) | System for monitoring disease progression | |
| Chen et al. | An Interpretable Deep Learning Optimized Wearable Daily Detection System for Parkinson’s Disease | |
| Li et al. | Evaluating focused attention level using mouse operation data | |
| Burt et al. | Essential tremor measurement and 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 |