CN107403061A - User's medical assessment model building method and medical assessment server - Google Patents
User's medical assessment model building method and medical assessment server Download PDFInfo
- Publication number
- CN107403061A CN107403061A CN201710549371.9A CN201710549371A CN107403061A CN 107403061 A CN107403061 A CN 107403061A CN 201710549371 A CN201710549371 A CN 201710549371A CN 107403061 A CN107403061 A CN 107403061A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- model
- sample
- medical assessment
- 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
Classifications
-
- 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
- A61B5/0205—Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
-
- 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
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- Physiology (AREA)
- Biophysics (AREA)
- Animal Behavior & Ethology (AREA)
- Cardiology (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Surgery (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Fuzzy Systems (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Mathematical Physics (AREA)
- Pulmonology (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The present invention, which provides a kind of user's medical assessment model building method and medical assessment server, the model, to be included：Step 101：Structure sample data U and sample input data V, the sample data U obtain sample input data V, V=the Ω .U through blurring mapping, and wherein Ω is the weight vectors of fuzzy quantization；Step 102：Clinical diagnosis result based on sample data U, build sample output result y, step 103：Establish neutral net mapping relations between sample input data V and sample output result y；Step 104：The neutral net of the model is trained；Inputted the sample input data V as the neutral net of the model, the sample output result y exports as the neutral net of the model；Learnt by sample training, determine the weights in the neutral net of the model.The user's medical assessment model building method and medical assessment server of the present invention, with reference to the gathered data of intelligent terminal, portable reliable medical diagnosis service can be provided the user.
Description
Technical field
The present invention relates to computer realm, more particularly to a kind of medical assessment model building method and medical assessment service
Device.
Background technology
As the improvement of people's living standards, irrational life and diet style, caused chronic disease and concurrent
Disease is just gradually affecting the health of people.In addition, aging aggravation, the aged's increases, old group health also by
To the influence of senile chronic disease.Medical science shows that the deterioration of these chronic diseases is can be by human body physiological data index
Monitoring and find in advance, so seeming very intentional using collection of the intelligent medical assessment system to physiological data and assessment
Justice.
Mainly there are two kinds of technical schemes on the structure of existing intelligent medical assessment system.It is single wear respectively
Wear equipment monitor system, and the assessment monitor system that wearable device is combined with equipment such as smart mobile phones.
The wearable device such as the first system solution, mainly intelligent postoperative care paster, intelligent watch or bracelet,
Its peripheral sensors gathers the physiology sign data of user, and shows on product screen.The cost of this technical scheme is relatively low, but
Medical assessment is not high, and accuracy is poor.
Several sensors on second of system solution, mainly intelligent watch or bracelet collect the physiology of user
Sign data, by bluetooth or wirelessly gathered data is sent in the equipment such as the smart mobile phone of user, on smart mobile phone
Using the result that each gathered data is assessed using simple Algorithm Analysis.User can call check recent gathered data and
Corresponding assessment result.This technical scheme improves medical assessment and analysis result accuracy, can also store a period of time
Data are used as reference data variation tendency.But comprehensive analysis assessment, computing capability can not be carried out to long-term substantial amounts of gathered data
Also it is not enough to carry out the data of collection information convergence analysis, assessment result lacks the accuracy of medical applications level.
Above scheme shows that current intelligent terminal can't be used for medical diagnosis, for chronic disease human patients or latent
Chronic disease patient groups, it is convenient to enjoy reliable medical services how with reference to intelligent terminal, not yet propose have at present
The solution of effect.
The content of the invention
The invention provides a kind of user's medical assessment model building method and medical assessment server, with reference to intelligent terminal
Gathered data, portable reliable medical diagnosis service can be provided the user.
The present invention provides a kind of user's medical assessment model building method, including：
Step 101：Building sample data U and sample input data V, U includes core signal ECG characteristic value data, blood oxygen
Data SpO2, heart rate data HR, blood pressure data BP, core signal ECG characteristic value data include：Phase, PR between phase, QRS wave between RR
Between phase, ST sections, R wave amplitudes, P wave amplitudes and T wave amplitudes between phase, QT；Sample data U obtains sample input data through blurring mapping
V, V=Ω .U, wherein Ω are the weight vectors of fuzzy quantization；
Step 102：Clinical diagnosis result based on sample data U, sample output result y is built,
Step 103：Establish neutral net mapping relations between sample input data V and sample output result y；
Step 104：The neutral net of model is trained；Sample input data V is defeated as the neutral net of model
Enter, sample output result y exports as the neutral net of model；Learnt by sample training, in the neutral net for determining model
Weights.
The present invention also provides a kind of medical assessment server, including：Including data module and user's medical assessment module.
Data module, periodic collection and the physiology sign data for preserving user terminal collection, physiology sign data at least wrap
Include core signal ECG, oximetry data SpO2, heart rate data HR and blood pressure data BP；
User's medical assessment module：The user preserved using user's medical assessment model of the present invention, analyze data module
Data, according to the output result of model, judge user's onset risk.
User's medical assessment model building method provided by the invention, the physiology sign number of user terminal collection can be combined
According to, export reliable medical diagnosis result, and the model can self study improve the accuracy of output result.So that using the medical treatment
The medical assessment server of assessment models, accurate, reliable, easily medical services can be provided the user.Exported by model
As a result automatic early-warning risk is seen a doctor or server proactive notification doctor provides medical services, instead of it is existing it is uncomfortable after again
Doctor, can allow patient at the initial stage of a disease or acute period of disease is treated in time, avoid delaying optimal treatment phase and delay treatment,
Reduce patient and treat cost, improve the experience sense of existing medical services and ageing.
Brief description of the drawings
Fig. 1 is the flow chart of user's medical assessment model building method of the present invention；
Fig. 2 is that user's medical assessment model of the present invention corrects flow chart in real time；
Fig. 3 is the structural representation of medical assessment server of the present invention.
Embodiment
In order that the object, technical solutions and advantages of the present invention are clearer, below in conjunction with the accompanying drawings with specific embodiment pair
The present invention is described in detail.
User's medical assessment model building method of the present invention, as shown in figure 1, comprising at least：
Step 101：Build sample data U and sample input data V.
Sample data U includes core signal ECG characteristic value data, oximetry data SpO2, heart rate data HR, blood pressure data
BP, core signal ECG characteristic value data include：Phase, ST sections, R wave amplitudes, P ripples between phase, QT between phase, PR between phase, QRS wave between RR
Amplitude and T wave amplitudes；It is fuzzy quantization that sample data U obtains sample input data V, V=Ω .U, wherein Ω through blurring mapping
Weight vectors.
Step 102：Clinical diagnosis result based on sample data U, structure sample output result y.
For example, the micro-judgment onset risk y of user's sample data can be defined as to normal, low-risk and excessive risk 3
Kind, y is used respectively0、y1、y2Represent, or further refine risk class.
Step 103：And establish neutral net mapping relations between sample input data V and sample output result y.
Step 104：The neural network model of medical assessment model is trained；Using sample input data V as user
The neutral net input of medical assessment model, sample output result y export as the neutral net of user's medical assessment model；It is logical
Sample training study is crossed, determines the weights in the neutral net of user's medical assessment model.
User's medical assessment model building method of the application, the sample data of each user can be directed to, train simultaneously structure
Different user's medical assessment models is built, because the physiology sign Index of each user differs from one another, same physiology sign refers to
Some numerical value of target, it is healthy numerical value for some users, and for other users, then it is dangerous numerical value, therefore adopt
With user's medical assessment model of the application, it is possible to achieve personalized customization so that the ratio of precision in general statistics mould of the model
Type is higher.
After user's medical assessment model being established based on step 101 to step 103, you can be used to assess in real time by model；Will
The user data U ' newly collected is obscured and is converted to V ', after the model being input to, model output result y', according to output result
Y' and sample output result y comparison, judge user's onset risk.
Further, user's medical model of the application, is also equipped with self-learning function, can the real-time assessment knot based on model
Fruit is constantly corrected, sophisticated model, as shown in Figure 2.
Modifying model：By the model output result y' of the user data newly collected compared with clinical diagnosis result,
If output result y' is not inconsistent with clinical diagnosis result, the numerical difference that output error is output result y' and diagnostic result is calculated
Value, output error is missed output to input layer successively anti-pass by the hidden layer of the neutral net of user's medical assessment model
Difference spreads out all units to each layer of medical assessment algorithm of neural network.
Fig. 3 is the application another embodiment, a kind of medical assessment server, including at least data module and medical assessment
Module.
Data module, periodic collection and the physiology sign data for preserving user terminal collection, physiology sign data at least wrap
Include core signal ECG, oximetry data SpO2, heart rate data HR and blood pressure data BP；
User's medical assessment module：Use user's medical assessment model of the application any embodiment, analyze data module
The user data of preservation, judge user's onset risk.
In order to improve the accuracy of data module stored data, before data preservation, first to the physiology of user's collection
Sign data is pre-processed, including the rejecting to abnormal data and non-user physiology sign data, to ensure to be input to user
The uniformity and stability of medical assessment model data.
The medical server of the application preserves user's physiology sign data, and packet is packaged into certain form, stores
In data module.Data packet format, encoded according to the form of packet header addend evidence, its middle wrapping head include start bit, device number,
Timestamp, physiology sign and status data name and check bit.Start bit is used to distinguish every segment data bag；Device number is aobvious for distinguishing
Show the distinct device of user；Timestamp is used to record acquisition time；Physiology sign data name is mainly used in distinguishing different sensors
The physiology sign data of collection and the different status information of user, electrocardiogram (ECG) data or heart rate number when such as distinguishing the data of collection
According to etc.；Check bit is used for whether verification data transmission to malfunction.
The medical assessment server of the application is additionally included in line doctor, when the user data that data module preserves is less,
Online doctor audits the output result of user's medical assessment model, and auditing result is fed back into user's medical assessment model, uses
In the real-time amendment of user's medical assessment model.
Or the data that online doctor preserves according to the output result and data module of user's medical assessment model, to user
Related medical diagnostic recommendations are provided, and diagnostic recommendations are sent to user terminal.When user's medical assessment model can be exported independently
During reliable judged result, the output result of user's medical assessment model can also be transmitted directly to user terminal by server.
If the output result of user's medical assessment model is " excessive risk ", server automatically turns on real-time monitoring module,
The real-time upload user physiology sign data of user terminal, server real-time monitoring user data, if monitoring unexpected abnormality, service
Device notifies online doctor to take medical measure at once, provides the user medical services.
Medical assessment model building method provided by the invention, the physiology sign data output of user terminal collection can be combined
Reliable medical diagnosis result, and the model can self study improve the accuracy of output result.So that using the medical assessment mould
The medical assessment server of type, accurate, reliable, easily medical services can be provided the user.By model output result certainly
Dynamic early warning risk is seen a doctor or server proactive notification doctor provides medical services, instead of it is existing it is uncomfortable after see a doctor again, can
To allow patient at the initial stage of a disease or acute period of disease is treated in time, avoid delaying optimal treatment phase and delay treatment, reduce
Patient treats cost, improves the experience sense of existing medical services and ageing.
The foregoing is merely illustrative of the preferred embodiments of the present invention, not to limit the present invention scope, it is all
Within the spirit and principle of technical solution of the present invention, any modification, equivalent substitution and improvements done etc., this hair should be included in
Within bright protection domain.
Claims (7)
1. a kind of user's medical assessment model building method, it is characterised in that the structure of the model comprises at least：
Step 101：Building sample data U and sample input data V, the U includes core signal ECG characteristic value data, blood oxygen
Data SpO2, heart rate data HR, blood pressure data BP, the core signal ECG characteristic value data include：Between RR between phase, QRS wave
Phase, ST sections, R wave amplitudes, P wave amplitudes and T wave amplitudes between phase, QT between phase, PR；The sample data U obtains institute through blurring mapping
It is the weight vectors of fuzzy quantization to state sample input data V, V=Ω .U, wherein Ω；
Step 102：Based on the clinical diagnosis result of the sample data U, sample output result y is built,
Step 103：Establish neutral net mapping relations between the sample input data V and the sample output result y；
Step 104：The neutral net of the model is trained；God using the sample input data V as the model
Through network inputs, the sample output result y exports as the neutral net of the model；Learnt by sample training, it is determined that
Weights in the neutral net of the model.
2. model according to claim 1, it is characterised in that the model also includes：
Modifying model：The user data U ' newly collected is converted into V ' through fuzzy, the institute being input to after step 103 training
State in model, output result y', if output result y' is not inconsistent with clinical diagnosis result, calculating output error is output result y'
With the quantity difference of the clinical diagnosis result, by the output error by the hidden layer of the neutral net of the model to defeated
Enter layer successively anti-pass, the output error is shared into all units to each layer of the neutral net of the model.
3. a kind of medical assessment server, it is characterised in that including at least data module and user's medical assessment module；
Data module, periodic collection and the physiology sign data for preserving user terminal collection, the physiology sign data at least wrap
Include core signal ECG, oximetry data SpO2, heart rate data HR and blood pressure data BP；
User's medical assessment module：Usage right requires any described user's medical assessment models of 1-2, analyzes the data mould
The user data that block preserves, according to the output result of the model, judges user's onset risk.
4. server according to claim 3, it is characterised in that when the user data that the data module preserves is less
When, the output result of user's medical assessment model is sent to doctor and audited by the server, and the doctor is audited
As a result user's medical assessment model is fed back to.
5. server according to claim 3, it is characterised in that the server is by user's medical assessment model
The data that output result and the data module preserve are sent to diagnosis, provide the user with related medical diagnostic recommendations, and
Suggest sending by described；Or the server directly transmits the output result of user's medical assessment model.
6. server according to claim 3, it is characterised in that the server also includes real-time monitoring module,
When the output result of user's medical assessment model is " excessive risk ", the server automatically turns on real-time monitoring module, uses
Family terminal uploads user's physiology sign data, the server real-time monitoring user data in real time, if it is different to monitor burst
Often, the server is given notice at once.
7. server according to claim 3, it is characterised in that when the data module preserves data, with preset data
Bag form stores；
The preset data bag form is encoded by the form of packet header addend evidence, the packet header include start bit, device number, when
Between stamp, physiology sign and status data name and check bit；The start bit is used to distinguish every segment data bag；The device number is used for
Distinguish the distinct device for showing user；The timestamp is used to record acquisition time；The physiology sign data name is used to distinguish
The physiology sign data and the different status information of user of different sensors collection, the check bit are used for verification data transmission
Whether malfunction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710549371.9A CN107403061A (en) | 2017-07-07 | 2017-07-07 | User's medical assessment model building method and medical assessment server |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710549371.9A CN107403061A (en) | 2017-07-07 | 2017-07-07 | User's medical assessment model building method and medical assessment server |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107403061A true CN107403061A (en) | 2017-11-28 |
Family
ID=60404897
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710549371.9A Pending CN107403061A (en) | 2017-07-07 | 2017-07-07 | User's medical assessment model building method and medical assessment server |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107403061A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108095708A (en) * | 2018-01-19 | 2018-06-01 | 动析智能科技有限公司 | A kind of physiology monitoring and analysis method, system based on mixing sensing |
| CN109767836A (en) * | 2018-12-29 | 2019-05-17 | 上海亲看慧智能科技有限公司 | A kind of medical diagnosis artificial intelligence system, device and its self-teaching method |
| CN111317440A (en) * | 2018-12-13 | 2020-06-23 | 通用电气公司 | Early warning method for patient, monitoring device using the method and readable storage medium |
| CN111353972A (en) * | 2018-12-21 | 2020-06-30 | 财团法人工业技术研究院 | State evaluation system, diagnosis and treatment system and operation method thereof |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1477581A (en) * | 2003-07-01 | 2004-02-25 | �Ϻ���ͨ��ѧ | Predictive modelling method application to computer-aided medical diagnosis |
| CN103038772A (en) * | 2010-03-15 | 2013-04-10 | 新加坡保健服务集团有限公司 | Method of predicting the survivability of a patient |
| CN105956374A (en) * | 2016-04-21 | 2016-09-21 | 青岛大学 | Remote monitoring comprehensive medical system |
| CN106021941A (en) * | 2016-05-27 | 2016-10-12 | 成都信汇聚源科技有限公司 | Method for obtaining risk predictions of sudden death with weight value matrices of artificial neural network |
| CN106407644A (en) * | 2016-08-03 | 2017-02-15 | 无锡金世纪国民体质与健康研究有限公司 | Assessment system used for health risk assessment |
-
2017
- 2017-07-07 CN CN201710549371.9A patent/CN107403061A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1477581A (en) * | 2003-07-01 | 2004-02-25 | �Ϻ���ͨ��ѧ | Predictive modelling method application to computer-aided medical diagnosis |
| CN103038772A (en) * | 2010-03-15 | 2013-04-10 | 新加坡保健服务集团有限公司 | Method of predicting the survivability of a patient |
| CN105956374A (en) * | 2016-04-21 | 2016-09-21 | 青岛大学 | Remote monitoring comprehensive medical system |
| CN106021941A (en) * | 2016-05-27 | 2016-10-12 | 成都信汇聚源科技有限公司 | Method for obtaining risk predictions of sudden death with weight value matrices of artificial neural network |
| CN106407644A (en) * | 2016-08-03 | 2017-02-15 | 无锡金世纪国民体质与健康研究有限公司 | Assessment system used for health risk assessment |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108095708A (en) * | 2018-01-19 | 2018-06-01 | 动析智能科技有限公司 | A kind of physiology monitoring and analysis method, system based on mixing sensing |
| WO2019142119A1 (en) * | 2018-01-19 | 2019-07-25 | 动析智能科技有限公司 | Hybrid sensing based physiological monitoring and analysis method and system |
| CN111317440A (en) * | 2018-12-13 | 2020-06-23 | 通用电气公司 | Early warning method for patient, monitoring device using the method and readable storage medium |
| CN111353972A (en) * | 2018-12-21 | 2020-06-30 | 财团法人工业技术研究院 | State evaluation system, diagnosis and treatment system and operation method thereof |
| CN111353972B (en) * | 2018-12-21 | 2024-03-12 | 财团法人工业技术研究院 | State evaluation system, diagnosis and treatment system and operation method thereof |
| CN109767836A (en) * | 2018-12-29 | 2019-05-17 | 上海亲看慧智能科技有限公司 | A kind of medical diagnosis artificial intelligence system, device and its self-teaching method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP3238491U (en) | An intelligent system that adjusts the optimal rehab intensity or amount of exercise to match the individual's exercise prescription | |
| CN107403061A (en) | User's medical assessment model building method and medical assessment server | |
| Ertin et al. | AutoSense: unobtrusively wearable sensor suite for inferring the onset, causality, and consequences of stress in the field | |
| CN102883775B (en) | Intracardiac electrogram time frequency noise detection | |
| CN107169307A (en) | Health risk assessment method and apparatus | |
| CN108478209A (en) | Ecg information dynamic monitor method and dynamic monitor system | |
| KR101596662B1 (en) | Apparatus and the method for assessing treatment response in obstructive sleep apenea | |
| KR20180007006A (en) | Appraisal | |
| CN109411041A (en) | Ecg information processing method and electro cardio signal workstation system | |
| CN113080878A (en) | System and method for physiological parameter monitoring | |
| KR20200092204A (en) | Chronic heart failure monitoring system using chest tissue resistance value and using method thereof | |
| CN108577830A (en) | A kind of user oriented sign information dynamic monitor method and dynamic monitor system | |
| CN101006913A (en) | A method for collecting data of physiological index using wireless network | |
| US20180206733A1 (en) | Device, method and system for monitoring and management of changes in hemodynamic parameters | |
| CN109360613A (en) | Follow-up method, device, computer equipment and the storage medium of user health | |
| CN113573627A (en) | System for measuring heart rate | |
| CN107205654B (en) | Method and device for adjusting a monitoring system | |
| CN108618788A (en) | A kind of psychological pressure apparatus for evaluating and psychological pressure appraisal procedure | |
| CN105249958B (en) | Human body information monitoring method, apparatus and system | |
| EP4057288A1 (en) | Enhanced reporting and charting of vital signs and other patient parameters | |
| CN115910346A (en) | Monitoring and early warning system and method for diabetic patient | |
| CN107495947A (en) | Blood pressure dynamic analysis method and blood pressure measuring device | |
| US11850053B2 (en) | System for measuring heart rate | |
| CN108903956A (en) | It detects a lie device and method | |
| CN101009035A (en) | Method for collecting date of physiological indexes using wireless network |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171128 |