CN109330846B - Air wave pressure massage instrument parameter optimization method based on deep learning algorithm - Google Patents
Air wave pressure massage instrument parameter optimization method based on deep learning algorithm Download PDFInfo
- Publication number
- CN109330846B CN109330846B CN201811141039.XA CN201811141039A CN109330846B CN 109330846 B CN109330846 B CN 109330846B CN 201811141039 A CN201811141039 A CN 201811141039A CN 109330846 B CN109330846 B CN 109330846B
- Authority
- CN
- China
- Prior art keywords
- deep learning
- air wave
- massage instrument
- user
- training
- 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.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H9/00—Pneumatic or hydraulic massage
- A61H9/005—Pneumatic massage
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/021—Measuring pressure in heart or blood vessels
-
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H2201/00—Characteristics of apparatus not provided for in the preceding codes
- A61H2201/50—Control means thereof
- A61H2201/5007—Control means thereof computer controlled
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Animal Behavior & Ethology (AREA)
- Artificial Intelligence (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Pathology (AREA)
- Cardiology (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Physiology (AREA)
- Heart & Thoracic Surgery (AREA)
- Fuzzy Systems (AREA)
- Evolutionary Computation (AREA)
- Vascular Medicine (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychiatry (AREA)
- Signal Processing (AREA)
- Epidemiology (AREA)
- Pain & Pain Management (AREA)
- Physical Education & Sports Medicine (AREA)
- Rehabilitation Therapy (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
- Massaging Devices (AREA)
Abstract
The invention discloses an air wave pressure massage instrument parameter optimization method based on a deep learning algorithm, and relates to an air wave pressure massage instrument parameter optimization method. The invention aims to solve the problems of low accuracy and high retraining cost caused by only once training of the existing deep learning. The invention comprises the following steps: firstly, the method comprises the following steps: collecting blood pressure and pulse data of a user; II, secondly: establishing a deep learning structure model on a server; thirdly, the method comprises the following steps: obtaining a trained deep learning structure model; fourthly, the method comprises the following steps: collecting blood pressure and pulse data of a user, inputting the blood pressure and pulse data into the trained deep learning structure model, and adjusting massage strength, massage frequency and massage parts of the air wave pressure massage instrument according to output air wave pressure massage instrument parameters; fifthly: training the obtained deep learning structure model after training through time T to obtain a new model; and (5) replacing the deep learning structure model trained in the step four with a new model, and repeatedly executing the step four. The invention is used in the technical field of medical treatment.
Description
Technical Field
The invention relates to the technical field of medical treatment, in particular to an air wave pressure massage instrument parameter optimization method based on a deep learning algorithm.
Background
The air wave pressure therapeutic apparatus forms circulation pressure to limbs and tissues by sequentially and repeatedly inflating and deflating the multi-cavity air bags, can promote the flow of blood and lymph, and is widely applied to treating or relieving various diseases. However, the following problems still remain:
(1) most of the massagers in the market can only complete simple massage operation and can only monitor the working state of the massagers.
(2) When data analysis is carried out on human health by the existing medical products, the existing medical products can only be generally used for analyzing a certain disease, and the existing medical products are expensive and cannot be borne by ordinary people.
(3) In the existing medical big data prediction, parameter setting is often dependent on empirical values and is difficult to maintain and adapt to changes.
(4) Most of the existing machine learning algorithms rely on existing empirical data, and are difficult to adapt to changes quickly after learning is completed.
In addition, artificial deep learning is a research hotspot which is raised in the field of artificial intelligence since the 80 s of the 20 th century. Deep learning is an operational model, which simulates the information processing mode of human brain neurons and sets a large number of information nodes (neurons), each node representing a specific output function called an excitation function. The connection between each two nodes represents a weighted value, called weight, for the signal passing through the connection, which is the basis for machine learning through artificial networks. The artificial deep learning is formed by mutually connecting a large number of nodes through weights, and different networks are formed among the nodes according to different connection modes. It is also often directly referred to in the engineering and academic circles as deep learning or quasi-deep learning. The output of the network is different according to the connection mode of the network, the weight value and the excitation function. The network itself is usually an approximation to some algorithm or function in nature, and may also be an expression of a logic strategy.
At present, big data technology and deep learning algorithm have been applied in many fields of medicine, such as diagnosis of special diseases, pathological research, disease prevention, etc. In addition, computer-aided diagnosis systems are gradually arranged in many high-grade hospitals abroad and some top-grade hospitals in China, so that the workload and the misdiagnosis rate of doctors are effectively reduced, and the medical efficiency is greatly improved.
Disclosure of Invention
The invention aims to solve the defects of low accuracy and high retraining cost caused by only once training of the existing deep learning, and provides an air wave pressure massage instrument parameter optimization method based on a deep learning algorithm.
An air wave pressure massage instrument parameter optimization method based on a deep learning algorithm comprises the following steps:
the method comprises the following steps: collecting blood pressure and pulse data of a user as a training set;
step two: establishing a deep learning structure model on a server;
step three: inputting the training set in the first step into the deep learning structure model established in the second step for training to obtain a trained deep learning structure model;
step four: collecting blood pressure and pulse data of a user of the air wave pressure massage instrument, inputting the blood pressure and pulse data into a trained deep learning structure model, outputting parameters of the air wave pressure massage instrument by the model, and adjusting massage force, massage frequency and massage part of the air wave pressure massage instrument according to the output parameters of the air wave pressure massage instrument, wherein the parameters of the air wave pressure massage instrument comprise the massage force, the massage frequency and the massage part; meanwhile, the server stores the collected blood pressure and pulse data of the air wave pressure massage instrument user and the air wave pressure massage instrument parameters output by the model;
step five: training the obtained deep learning structure model after training by adopting the blood pressure and pulse data of the air wave pressure massage instrument user stored by the server in the step four and the air wave pressure massage instrument parameters output by the model after the time T to obtain a new model; and (5) replacing the deep learning structure model trained in the step four with a new model, and repeatedly executing the step four.
The invention has the beneficial effects that:
on the basis of the air wave pressure instrument, the invention adds a part for detecting various health data indexes of the human body, can monitor the data of the blood pressure, the pulse and the like of the human body while finishing massage and detecting the working state of the human body, and adjusts the working state according to various indexes of the human body, thereby bringing more healthy and comfortable enjoyment to people; meanwhile, the traditional deep learning algorithm is improved: compared with the traditional machine learning, the algorithm adopted by the invention only adopts data in a period of time as training data, and except for the first machine learning, the last parameter value is adopted as the initial parameter value of the next training, and when the result obtained continuously for multiple times is not consistent with the actual result, the fine adjustment is carried out by taking the original data as the center, and a new round of learning is carried out, so that the machine can learn more quickly and can adapt to the change; based on the improved algorithm, performing multi-data analysis on various data indexes of the human body by utilizing big data (historical data of the user and historical data of a large number of other users), calculating the correlation between various data and the health condition of the human body, continuously adjusting correction parameters in the process, and providing health prediction for the user and making the most appropriate health care scheme; the purpose of treating the disease before existence is achieved.
The measuring equipment is combined with the physiotherapy equipment, so that the trouble of redundant and complicated equipment is avoided.
Compared with the existing medical big data service, the method and the system have the advantages that the assisted population is better diverted from doctors to users, and the users can more conveniently benefit from big data analysis.
The health data detection of the user and the adjustment of the working state of the instrument are carried out in real time, so that the disease prevention capability and the response speed are improved, and the purpose of treating diseases and not treating diseases is conveniently achieved.
The training data is timeliness, compared with the traditional deep learning algorithm, the method can quickly and accurately adapt to the change of various body indexes of the user, meanwhile, parameter adjustment during multiple times of training is not completely retrained but fine adjustment is carried out on the original basis, and the retraining cost is reduced.
The invention saves the time of repeated training: the training time of the traditional deep learning generally needs dozens of hours to hundreds of hours, while the method adopted by the invention only needs two hours except the first training.
The invention applies the deep learning algorithm to the air wave massage instrument for the first time and fills the market blank.
In one month of trial use by volunteers, the air wave massage instrument of the invention is recorded to change different working states for 16 times, compared with the common air wave massage instrument with only a plurality of gears on the market, the air wave massage instrument greatly improves the flexibility and diversity of massage modes; after one month, the data of the volunteers in one month is taken as the training set for retraining for only 1 hour and 53 minutes, the accuracy is 96.7 percent, and the retraining cost is greatly reduced.
Drawings
Fig. 1 is a schematic working flow diagram of an air wave pressure massage apparatus.
Detailed Description
The first embodiment is as follows: an air wave pressure massage instrument parameter optimization method based on a deep learning algorithm comprises the following steps:
the method comprises the following steps: collecting blood pressure and pulse data of a user as a training set;
step two: establishing a deep learning structure model on a server;
step three: inputting the training set in the first step into the deep learning structure model established in the second step for training to obtain a trained deep learning structure model;
step four: collecting blood pressure and pulse data of a user of the air wave pressure massage instrument, inputting the blood pressure and pulse data into a trained deep learning structure model, outputting parameters of the air wave pressure massage instrument by the model, and adjusting massage force, massage frequency and massage part of the air wave pressure massage instrument according to the output parameters of the air wave pressure massage instrument, wherein the parameters of the air wave pressure massage instrument comprise the massage force, the massage frequency and the massage part; meanwhile, the server stores the collected blood pressure and pulse data of the air wave pressure massage instrument user and the air wave pressure massage instrument parameters output by the model;
step five: training the obtained deep learning structure model after training by adopting the blood pressure and pulse data of the air wave pressure massage instrument user stored by the server in the step four and the air wave pressure massage instrument parameters output by the model after the time T to obtain a new model; and (5) replacing the deep learning structure model trained in the step four with a new model, and repeatedly executing the step four.
An intelligent massage physiotherapy instrument based on an air wave treatment method comprises a massage air cushion, an air bag and a central control circuit (comprising a main control chip, a pressure sensor, a Bluetooth module, a buzzer and a filtering and amplifying circuit), and the working process of the intelligent massage physiotherapy instrument is shown in figure 1.
The pressure sensor of the air wave massager acquires data such as blood pressure, pulse and the like of a user, and the Bluetooth module of the air wave massager transmits the data acquired by the pressure sensor to the mobile phone; the mobile phone uploads the data acquired by the air wave massager to the server through the matched APP; the server judges the body state of the user at the moment according to the received data acquired by the air wave massager; the server selects the massage frequency, the massage force and the massage part of the air wave massage instrument according to the calculated body state of the user, and sends an adjusting instruction to the APP matched with the mobile phone; the APP matched with the mobile phone sends the instruction to the air wave massage instrument through the Bluetooth after receiving the adjustment instruction of the server, and the air wave massage instrument is controlled to change massage frequency, force and position.
The air wave physiotherapy instrument provides massage service, transmits data of heartbeat, blood pressure and the like of a user and the working state of the user to a mobile phone application through Bluetooth communication through a pressure sensor and a filtering and amplifying circuit, uploads the data to a cloud server after application, judges the state of the user at the moment and adjusts parameters after the server calculates through a deep learning algorithm, and sends the state to the mobile phone application control physiotherapy instrument to change the working state.
The cloud server part follows the deep learning algorithm, and various conditions of deep learning are as follows:
the excitation function is set as follows: according to the existing international normal standard range, if the user value is higher than the range, the user value is marked as the upper boundary of the user value-range, otherwise, the user value is marked as the lower boundary of the user value-range. Then taking the softplus function as the excitation function.
The network structure is as follows: and a two-layer deep learning framework is adopted.
Training data sources: large medical development source data sets and user historical data at home and abroad.
In the initial starting state, the number of times that the calculation result does not accord with the fact is set as the error function by giving random values of the average distribution between 0 and 1 with different weight functions and utilizing the existing starting data set (namely the fact is that the user is healthy, the calculation result is that the user is unhealthy, or the opposite is true). After sufficient training, a set of weight functions with the smallest error function is selected as the initial weight values.
After the full training, the determined model is uploaded to a cloud server, and the sent user data is processed. If the calculated result of the numerical substitution of the user continuously exceeds the preset threshold value for a period of time, the health of the user at the moment is judged to be risky, and the condition which has the largest influence in the final result is detected, so that the working state (strength, frequency, position and the like) of the machine is adjusted according to the condition.
Meanwhile, every time, the deep learning algorithm retrains the user data acquired in the time by taking the last parameter value as the center once so as to adapt to the change of the body state of the user more quickly and better
In addition, the deep learning algorithm can arrange historical data of a user after the user uses the historical data for a period of time, calculate the average value, the change rate, the variance and the like of each index of the user in the period of time, and analyze whether the trend of the user deviates from a normal value or is kept in a normal state according to the average value, the change rate, the variance and the like; based on the above, the health condition analysis and suggestion are given to the user.
In one month of trial use by volunteers, the air wave massage instrument of the invention is recorded to change different working states for 16 times, compared with the common air wave massage instrument with only a plurality of gears on the market, the air wave massage instrument greatly improves the flexibility and diversity of massage modes; after one month, the data of the volunteers in one month is taken as the training set for retraining for only 1 hour and 53 minutes, the accuracy is 96.7 percent, and the retraining cost is greatly reduced.
The second embodiment is as follows: the first difference between the present embodiment and the specific embodiment is: the specific process of establishing the deep learning structure model on the server in the step two is as follows:
and (4) taking the blood pressure and pulse data of the user collected in the step one as input, and outputting the parameters of the air wave pressure massage instrument through the excitation function layer and the full connection layer.
Other steps and parameters are the same as those in the first embodiment.
The third concrete implementation mode: the present embodiment differs from the first or second embodiment in that: the excitation function of the excitation function layer is specifically as follows:
preprocessing the collected blood pressure and pulse data of the user;
setting threshold ranges of blood pressure and pulse; if the difference value is higher than the threshold range, the difference value between the upper limit value and the threshold range is marked as positive; if the difference value is lower than the threshold range, the difference value between the upper limit and the lower limit of the threshold is marked as negative; is marked as 0 in the threshold value range;
taking a softplus function as an excitation function for the preprocessed value;
the softplus function is log (1+ exp (t)), where t is the value after preprocessing.
Other steps and parameters are the same as those in the first or second embodiment.
The fourth concrete implementation mode: the difference between this embodiment mode and one of the first to third embodiment modes is: inputting the training set in the first step into the deep learning structure model established in the second step for training in the third step, wherein the specific process for obtaining the trained deep learning structure model is as follows:
by giving random values of average distribution between 0 and 1 with different weight functions, setting an error function as the times of the calculation result of blood pressure and pulse data of a user passing through a deep learning model and the health condition of the user, namely the times of the fact that the calculation result of the health of the user is unhealthy or the times of the fact that the calculation result of the unhealthy user is healthy, by using an existing open source data set as a training set; the specific training method is a layer-by-layer training method, namely, training one layer first, and after the training of the layer is finished, training the next layer by taking the output value of the layer as the input of the next layer until the training of each layer is finished. After sufficient training, a set of weight functions with the smallest error function is selected as the initial weight values.
Other steps and parameters are the same as those in one of the first to third embodiments.
The fifth concrete implementation mode: the difference between this embodiment and one of the first to fourth embodiments is: the fourth step is to collect the blood pressure and pulse data of the air wave pressure massage instrument user, input the data into the deep learning structure model after training, output the parameters of the air wave pressure massage instrument by the model, and adjust the massage force, massage frequency and massage part of the air wave pressure massage instrument according to the output parameters of the air wave pressure massage instrument, wherein the concrete process comprises the following steps:
after the deep learning structure model is fully trained, the determined model is uploaded to a cloud server, a massager collects blood pressure and pulse data of a user and sends the blood pressure and pulse data to the server, the server processes the blood pressure and pulse data sent by the massager, if the blood pressure and pulse data of the user input the result calculated by the deep learning model exceeds a preset threshold value for 3 consecutive days, the health of the user at the moment is judged to be risky, input data causing health risks are determined according to the maximum value of 4 detected data in the hidden layer 2, and a corresponding working mode of the massager is selected according to the determined input data causing health risks. Such as: the massage frequency, the massage force and the initial position are respectively set to be 5 seconds/time, 15kPa and the upper part of the shank; the blood pressure of the user is collected as follows: 130mmHg at high pressure and 90mmHg at low pressure; pulse rate is 62 times; the result obtained according to the deep learning algorithm is-0.1524, the pulse is not in the set threshold range (0.4500-0.7200), the influence of the pulse is detected to be large, and the massage frequency, the force and the position are respectively adjusted to be 3 seconds/time, 20kPa and all legs, so that the effect of accelerating the blood circulation is achieved.
Other steps and parameters are the same as in one of the first to fourth embodiments.
The sixth specific implementation mode: the difference between this embodiment and one of the first to fifth embodiments is: and step five, training the obtained deep learning structure model after training by adopting the blood pressure and pulse data of the air wave pressure massage instrument user stored by the server in step four and the air wave pressure massage instrument parameters output by the model after the time T passes, wherein the specific process of obtaining a new model is as follows:
based on the weight function obtained by the previous training, the weight function is given as an expectation obeying the weight function obtained by the previous training, 0.001 is a normally distributed random number with variance, the blood pressure and pulse data of the air wave pressure massage instrument user stored by the server in the T step four and the air wave pressure massage instrument parameter output by the model are used as a training set, and an error function is set as the times that the calculation result of the blood pressure and pulse data of the user passing through the deep learning model is inconsistent with the health condition of the user, namely the times that the user health calculation result is the unhealthy condition of the user or the times that the user unhealthy calculation result is the healthy condition of the user; after sufficient training, a set of weight functions with the smallest error function is selected as the initial weight values.
Other steps and parameters are the same as in one of the first to fourth embodiments.
The present invention is capable of other embodiments and its several details are capable of modifications in various obvious respects, all without departing from the spirit and scope of the present invention.
Claims (6)
1. A method for optimizing parameters of an air wave pressure massage instrument based on a deep learning algorithm is characterized by comprising the following steps: the air wave pressure massage instrument parameter optimization method based on the deep learning algorithm comprises the following steps:
the method comprises the following steps: collecting blood pressure and pulse data of a user as a training set;
step two: establishing a deep learning structure model on a server;
step three: inputting the training set in the first step into the deep learning structure model established in the second step for training to obtain a trained deep learning structure model;
step four: collecting blood pressure and pulse data of a user of the air wave pressure massage instrument, inputting the blood pressure and pulse data into a trained deep learning structure model, outputting parameters of the air wave pressure massage instrument by the model, and adjusting massage force, massage frequency and massage part of the air wave pressure massage instrument according to the output parameters of the air wave pressure massage instrument, wherein the parameters of the air wave pressure massage instrument comprise the massage force, the massage frequency and the massage part; meanwhile, the server stores the collected blood pressure and pulse data of the air wave pressure massage instrument user and the air wave pressure massage instrument parameters output by the model;
step five: training the obtained deep learning structure model after training by adopting the blood pressure and pulse data of the air wave pressure massage instrument user stored by the server in the step four and the air wave pressure massage instrument parameters output by the model after the time T to obtain a new model; and (5) replacing the deep learning structure model trained in the step four with a new model, and repeatedly executing the step four.
2. The air wave pressure massage instrument parameter optimization method based on the deep learning algorithm is characterized by comprising the following steps of: the specific process of establishing the deep learning structure model on the server in the step two is as follows:
and (4) taking the blood pressure and pulse data of the user collected in the step one as input, and outputting the parameters of the air wave pressure massage instrument through the excitation function layer and the full connection layer.
3. The air wave pressure massage instrument parameter optimization method based on the deep learning algorithm is characterized by comprising the following steps of: the excitation function of the excitation function layer is specifically as follows:
preprocessing the collected blood pressure and pulse data of the user;
setting threshold ranges of blood pressure and pulse; if the difference value is higher than the threshold range, the difference value between the upper limit value and the threshold range is marked as positive; if the difference value is lower than the threshold range, the difference value between the upper limit and the lower limit of the threshold is marked as negative; is marked as 0 in the threshold value range;
taking a softplus function as an excitation function for the preprocessed value;
the softplus function is log (1+ exp (t)), where t is the value after preprocessing.
4. The air wave pressure massage instrument parameter optimization method based on the deep learning algorithm is characterized by comprising the following steps of: inputting the training set in the first step into the deep learning structure model established in the second step for training in the third step, wherein the specific process for obtaining the trained deep learning structure model is as follows:
by giving random values of average distribution between 0 and 1 with different weight functions, setting an error function as the times of the calculation result of blood pressure and pulse data of a user passing through a deep learning model and the health condition of the user, namely the times of the fact that the calculation result of the health of the user is unhealthy or the times of the fact that the calculation result of the unhealthy user is healthy, by using an existing open source data set as a training set; the specific training method is a layer-by-layer training method, namely, training one layer first, and after the training of the layer is finished, training the next layer by taking the output value of the layer as the input of the next layer until the training of each layer is finished; after each layer is trained, a group of weight functions with the minimum error function is selected as initial weight values.
5. The air wave pressure massage instrument parameter optimization method based on the deep learning algorithm is characterized by comprising the following steps of: the fourth step is to collect the blood pressure and pulse data of the air wave pressure massage instrument user, input the data into the deep learning structure model after training, output the parameters of the air wave pressure massage instrument by the model, and adjust the massage force, massage frequency and massage part of the air wave pressure massage instrument according to the output parameters of the air wave pressure massage instrument, wherein the concrete process comprises the following steps:
after the deep learning structure model is trained, the determined model is uploaded to a server, a massage instrument collects blood pressure and pulse data of a user and sends the blood pressure and pulse data to the server, the server processes the blood pressure and pulse data sent by the massage instrument, if the blood pressure and pulse data of the user input a result calculated by the deep learning model exceeds a preset threshold value for 3 consecutive days, the health of the user at the moment is judged to be risky, input data causing health risks are determined according to the maximum value of 4 detected data in the hidden layer 2, and a corresponding working mode of the massage instrument is selected according to the determined input data causing health risks.
6. The air wave pressure massage instrument parameter optimization method based on the deep learning algorithm is characterized by comprising the following steps of: and step five, training the obtained deep learning structure model after training by adopting the blood pressure and pulse data of the air wave pressure massage instrument user stored by the server in step four and the air wave pressure massage instrument parameters output by the model after the time T passes, wherein the specific process of obtaining a new model is as follows:
giving a weight function as an expectation for the weight function obtained by following the last training, taking the normal distribution random number with the variance of 0.001, taking the blood pressure and pulse data of the air wave pressure massage instrument user stored by the server in the T step four and the parameters of the air wave pressure massage instrument output by the model as a training set, and setting an error function as the times that the calculation result of the blood pressure and pulse data of the user passing through the deep learning model is inconsistent with the health condition of the user, namely the times that the calculation result of the health of the user is unhealthy or the times that the calculation result of the unhealthy user is healthy; after training, a set of weight functions with the smallest error function is selected as the initial weight values.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811141039.XA CN109330846B (en) | 2018-09-28 | 2018-09-28 | Air wave pressure massage instrument parameter optimization method based on deep learning algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811141039.XA CN109330846B (en) | 2018-09-28 | 2018-09-28 | Air wave pressure massage instrument parameter optimization method based on deep learning algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109330846A CN109330846A (en) | 2019-02-15 |
| CN109330846B true CN109330846B (en) | 2021-03-12 |
Family
ID=65307290
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811141039.XA Active CN109330846B (en) | 2018-09-28 | 2018-09-28 | Air wave pressure massage instrument parameter optimization method based on deep learning algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109330846B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112206140B (en) * | 2020-09-03 | 2022-04-15 | 国家康复辅具研究中心 | Pneumatic circulation training system of heart brain coupling direction |
| CN112155979B (en) * | 2020-10-21 | 2021-08-31 | 北京航空航天大学 | Inflatable pressurization treatment equipment and optimal parameter acquisition method thereof |
| CN113116692B (en) * | 2021-04-19 | 2022-06-07 | 南京智能高端装备产业研究院有限公司 | Self-adaptive intermittent inflation pressurization system based on interface pressure |
| CN113499232A (en) * | 2021-06-03 | 2021-10-15 | 山西三友和智慧信息技术股份有限公司 | Remote massage system based on artificial intelligence |
| CN113712794B (en) * | 2021-08-23 | 2024-04-09 | 河北华蓝医疗器械有限公司 | Medical pressurization pressure monitoring system |
| CN113838578A (en) * | 2021-09-27 | 2021-12-24 | 南方医科大学珠江医院 | Big data-based trachea foreign matter emergency treatment device with striking hammer |
| CN114569423B (en) * | 2022-02-17 | 2023-11-21 | 湖南抖医供应链有限公司 | Digital intelligent sharing medical anti-thrombus robot |
| CN116909517B (en) * | 2023-09-11 | 2023-11-21 | 深圳市德壹医疗科技有限公司 | Massage robot massage program acquisition method and device based on big data |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6658396B1 (en) * | 1999-11-29 | 2003-12-02 | Tang Sharon S | Neural network drug dosage estimation |
| CN102525795A (en) * | 2012-01-16 | 2012-07-04 | 沈阳理工大学 | Fast automatic positioning method of foot massaging robot |
| CN104915560A (en) * | 2015-06-11 | 2015-09-16 | 万达信息股份有限公司 | Method for disease diagnosis and treatment scheme based on generalized neural network clustering |
| CN205015889U (en) * | 2015-09-23 | 2016-02-03 | 北京科技大学 | Definite system of traditional chinese medical science lingual diagnosis model based on convolution neuroid |
| CN106485054A (en) * | 2016-09-21 | 2017-03-08 | 广东工业大学 | Intelligent diagnostics data analysing method based on BP neural network algorithm and system |
| CN106580393A (en) * | 2016-11-24 | 2017-04-26 | 北京康易联技术有限公司 | Method and device for adjusting training parameter |
| CN106951719A (en) * | 2017-04-10 | 2017-07-14 | 荣科科技股份有限公司 | The construction method and constructing system of clinical diagnosis model, clinical diagnosing system |
| CN107194158A (en) * | 2017-05-04 | 2017-09-22 | 深圳美佳基因科技有限公司 | A kind of disease aided diagnosis method based on image recognition |
| CN107451407A (en) * | 2017-07-31 | 2017-12-08 | 广东工业大学 | A kind of traditional Chinese medical science intelligent diagnosing method, system and traditional medical system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150012471A1 (en) * | 2013-07-02 | 2015-01-08 | SenseGon Technologies Ltd. | System for user psychosocial profiling |
| US20170156592A1 (en) * | 2015-12-02 | 2017-06-08 | Mediatek Inc. | Healthcare systems and monitoring method for physiological signals |
-
2018
- 2018-09-28 CN CN201811141039.XA patent/CN109330846B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6658396B1 (en) * | 1999-11-29 | 2003-12-02 | Tang Sharon S | Neural network drug dosage estimation |
| CN102525795A (en) * | 2012-01-16 | 2012-07-04 | 沈阳理工大学 | Fast automatic positioning method of foot massaging robot |
| CN104915560A (en) * | 2015-06-11 | 2015-09-16 | 万达信息股份有限公司 | Method for disease diagnosis and treatment scheme based on generalized neural network clustering |
| CN205015889U (en) * | 2015-09-23 | 2016-02-03 | 北京科技大学 | Definite system of traditional chinese medical science lingual diagnosis model based on convolution neuroid |
| CN106485054A (en) * | 2016-09-21 | 2017-03-08 | 广东工业大学 | Intelligent diagnostics data analysing method based on BP neural network algorithm and system |
| CN106580393A (en) * | 2016-11-24 | 2017-04-26 | 北京康易联技术有限公司 | Method and device for adjusting training parameter |
| CN106951719A (en) * | 2017-04-10 | 2017-07-14 | 荣科科技股份有限公司 | The construction method and constructing system of clinical diagnosis model, clinical diagnosing system |
| CN107194158A (en) * | 2017-05-04 | 2017-09-22 | 深圳美佳基因科技有限公司 | A kind of disease aided diagnosis method based on image recognition |
| CN107451407A (en) * | 2017-07-31 | 2017-12-08 | 广东工业大学 | A kind of traditional Chinese medical science intelligent diagnosing method, system and traditional medical system |
Non-Patent Citations (1)
| Title |
|---|
| 遗传算法优化BP神经网络在中医按摩机器人中的应用;张秋云等;《应用科技》;20170430;第44卷(第2期);第73-77页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109330846A (en) | 2019-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109330846B (en) | Air wave pressure massage instrument parameter optimization method based on deep learning algorithm | |
| CN108498089B (en) | Noninvasive continuous blood pressure measuring method based on deep neural network | |
| CN107495962A (en) | A kind of automatic method by stages of sleep of single lead brain electricity | |
| CN107680683A (en) | A kind of AI eye healths appraisal procedure | |
| KR102424629B1 (en) | System and method for improving development disorder using deep learning module | |
| CN108814565A (en) | A kind of intelligent Chinese medicine health detection dressing table based on multi-sensor information fusion and deep learning | |
| CN106372729B (en) | Deep learning method and device for psychological analysis | |
| CN111899861A (en) | Intelligent nursing method and system for intensive care unit | |
| CN102058453B (en) | Electrical stimulation audition tone perception testing and estimating system | |
| CN110752032A (en) | Traditional Chinese medicine diagnosis method based on convolutional neural network and laser vibration measurement | |
| CN109599178A (en) | Threshold determination model determines method, apparatus, medical treatment detection device and storage medium | |
| CN117814766B (en) | Near infrared spectrum physical and mental pressure and sleep quality monitoring system | |
| CN118044813B (en) | Psychological health condition assessment method and system based on multitask learning | |
| CN115409764A (en) | Multi-mode fundus blood vessel segmentation method and device based on domain self-adaptation | |
| Zhang et al. | A hybrid model for blood pressure prediction from a PPG signal based on MIV and GA-BP neural network | |
| Chen et al. | A low cost, adaptive mixed reality system for home-based stroke rehabilitation | |
| CN104484562A (en) | Fuzzy neural network-based body area network health information monitoring and interacting system | |
| CN113040738B (en) | Blood pressure detecting device | |
| KR20220089913A (en) | System and method for improving development disorder using deep learning module | |
| Kumar et al. | Analysis of CNN model based classification of diabetic retinopathy diagnosis | |
| Shahraki et al. | The Impact of Brain–Computer Interface on Lifestyle of Elderly People | |
| CN114098661B (en) | Lower limb arteriosclerosis disease treatment system based on external counterpulsation technology | |
| CN108875580A (en) | A kind of multiclass Mental imagery EEG signal identification method based on because imitating network | |
| CN205814348U (en) | Cardiac valve cardiechema signals mechanical vibration intelligent analysis system | |
| CN113208569B (en) | Pulse wave curve fitting method based on group algorithm |
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 |