CN105760704A - Establishing method of angiosclerosis characteristic spectrum multi-parameter medical model and software system of method - Google Patents

Abstract

The invention discloses an establishing method of an angiosclerosis characteristic spectrum multi-parameter medical model and a software system of the method.The method includes the steps of calling multi-parameter dynamic hidden information of the inner wall and outer wall of a blood vessel from a database, analyzing each type of parameter to obtain a characteristic spectrum of the parameters of the walls of the blood vessel, obtaining power spectrum estimation of each type of parameter according to the corresponding characteristic spectrum, obtaining the analysis result of the multi-parameter dynamic hidden information according to a medical model, and establishing the software system on the basis of the angiosclerosis characteristic spectrum multi-parameter medical model.The method and the system have the advantages of being simple in step, accurate in calculation, high in accuracy and low in cost; the database for monitoring the health degree of people can be established according to the model result, calculated through the model, of healthy people and typical blood vessel disease patients, the data acquisition amount is large, and the database is a large database.

Description

The method for building up of sclerosis of blood vessels characteristic spectrum multiparameter medical model and software system thereof

Technical field

The present invention relates to medical model field, particularly relate to method for building up and the software system thereof of sclerosis of blood vessels characteristic spectrum multiparameter medical model.

Background technology

The quantitatively change of earlier acquisition vessel wall elasticity function is one of difficult problem in current genus of stiff vessel diseases clinical diagnosis, but, clinical medicine is still extremely limited to the diagnostic method of genus of stiff vessel typical disease so far, currently mainly there are some Morphologic Diagnosis methods, the blood vessel collection of such as doppler ultrasound, CT and MRI and radionuclide collection etc., not only its acquisition method is complicated, and diagnostic accuracy is relatively low, the overwhelming majority can only reflect local patholoic change and want PD just can collect to when having an obvious pathological change.

Summary of the invention

For solving above technical problem, the present invention provides method for building up and the software system of a kind of sclerosis of blood vessels characteristic spectrum multiparameter medical model, computer technology is utilized to detect eight noinvasive multidate informations, then the characteristic spectrum of these eight information is utilized, create sclerosis of blood vessels characteristic spectrum multiparameter medical model, as the judgement factor of prevention cardiovascular disease.

Technical scheme is as follows:

A kind of method for building up of sclerosis of blood vessels characteristic spectrum multiparameter medical model, it it is critical only that: method for building up comprises the following steps:

Step 1: transfer the dynamic implicit information of multiparameter inside and outside blood vessel wall from data base, the dynamic implicit information of described multiparameter includes the acceleration wave of blood vessel wall line pressure shift ripple, blood vessel wall line pressure shift wave velocity ripple, blood vessel wall pressure wave acceleration wave, cardiechema signals, electrocardio ripple, capacity of blood vessel ripple, the differential ripple of capacity of blood vessel ripple, capacity of blood vessel ripple；

Wherein electrocardio ripple is the reference signal gathered, blood vessel wall line pressure shift ripple, blood vessel wall line pressure shift wave velocity ripple, blood vessel wall pressure wave acceleration wave are the external dynamic information of blood vessel wall, and cardiechema signals, capacity of blood vessel ripple, the differential ripple of capacity of blood vessel ripple, cardiechema signals are blood vessel wall internal dynamic information；

Step 2: the signal of dynamic for the multiparameter collected implicit information is transformed from the time domain to frequency domain；

Step 3: every kind of dynamic state of parameters implicit information through frequency domain transform is analyzed；

Step 4: the characteristic spectrum of every kind of parameter is carried out power Spectral Estimation, sets up the characteristic spectrum power Spectral Estimation p of every kind of parameter_PER(w)；

Step 5: obtaining testing result according to the power Spectral Estimation of every kind of parameter and sclerosis of blood vessels characteristic spectrum multiparameter medical model, the formula of this multiparameter medical model is:

$H=\underset{i=1}{\overset{8}{\Σ}}{k}_i{P}_{PER\left(w\right)}$

Wherein, k_iRepresent the weighter factor of every kind of parameter, p_PER(w)Represent the characteristic spectrum power Spectral Estimation of every kind of parameter；

Adopt said method, by the multiparameter multidate information inside and outside the blood vessel wall of collection in analytical database, the multiparameter medical model result being able to multidate information inside and outside the blood vessel wall of multiparameter medical model and the sufferer crowd obtaining multidate information inside and outside the blood vessel wall of normal population is calculated in conjunction with sclerosis of blood vessels characteristic spectrum multiparameter medical model, two results are compared, obtain comparative result, comparative result is created as data base, and this data base just can be used for the big data monitoring of population health degree.

Further, step 2 adopt fast Fourier transform that the dynamic implicit information of multiparameter gathered carries out the conversion of time domain and frequency domain in conjunction with the method for temporal decimation.

Adopting said method, Fourier transformation is to use the widest converter technique, can meet the conversion from time domain to frequency domain of the blood vessel wall multiparameter multidate information, it is easy to differentiation and the process to signal.

Further, step 3 adopts following steps each parameter is analyzed:

Step 1., every kind of parameter is carried out trend analysis, eliminates the irregular variation in every kind of parameter, sets up mathematical model of the computer；

Step 2., adopt below equation every kind of parameter is carried out correlation analysis, obtain the universal law of disease

$r_{A,B}=\frac{\underset{t=1}{\overset{T}{\Σ}}(a_t-\stackrel{\‾}{A})(b_t-\stackrel{\‾}{B})}{{\mathrm{T\σ}}_A{\mathrm{\σ}}_B}$

Wherein, r_A,BRepresent A, B two relative coefficient of track data；a_t,b_tRepresent the t moment A, B two data value of track data respectively；T represents total time series number；Represent the average of data A and data B respectively；σ_A,σ_BRepresent the standard variance of data A and data B respectively, if r_A,BEqual to 0, then data A and data B is uncorrelated, otherwise is then correlated with.

Adopt above-mentioned steps, irregular deviation in the supplemental characteristic gathered can be eliminated；Obtain the dependency between different parameters data.

Further, adopting period map method that the characteristic spectrum of every kind of parameter of the dynamic implicit information of the multiparameter being transformed into frequency domain is carried out power Spectral Estimation in step 4, its formula is:

$P_{PER}\left(w\right)=\frac{1}{N}{\left|\underset{n=0}{\overset{n-1}{\Σ}}x\left(n\right)e^{-j\mathrm{\π}wn}\right|}^2=\frac{1}{N}{\left|X_N\left(w\right)\right|}^2$

X_NW () represents the frequency domain sequence of every kind of parameter；X (n) represents the time domain sequences of every kind of parameter, P_PERW () represents that the power of the characteristic spectrum of every kind of parameter is estimated, N represents the observation station quantity of every kind of input parameter signal.Adopting said method, period map method is widely used, and can guarantee that precision of analysis.

Further, 1. to adopt the method for rolling average to carry out its mathematic(al) representation of trend analysis as follows for step:

$\frac{t_1+t_2+\mathrm{...}+t_n}{n},\frac{t_2+t_3+\mathrm{...}+t_{n+1}}{n},\frac{t_3+t_4+\mathrm{...}+t_{n+2}}{n},\frac{t_4+t_5+\mathrm{...}+t_{n+3}}{n}$

Wherein t_nRepresenting the data value in the n-th moment, n represents the data bulk of collection.

Adopt said method, irregular deviation in time series data can be eliminated.

Further, a kind of software system based on sclerosis of blood vessels characteristic spectrum multiparameter medical model, it it is critical only that: includes data acquisition module, case management module, signal processing module, system management module and data memory module, the dynamic implicit information of blood vessel wall multiparameter of described data collecting module collected patient and patient data, and deliver the information to signal processing module and be analyzed, data acquisition module and signal processing module send case management module to patient data and analysis result respectively and form individual data, data memory module stores with individual data gathering data, system is configured and manages by system management module；

Described signal processing module is made up of signal filtering module and signal analysis module, the signal gathered is filtered by described signal filtering module, and described signal analysis module utilizes hardening characteristics spectrum multiparameter medical model that the multiparameter multidate information after after filtering is analyzed.

Adopt said system, can quickly process substantial amounts of supplemental characteristic, obtain the oscillogram of characteristic spectrum.

Further, described signal filtering module adopts following methods that signal is filtered:

Assuming that desired parameters signal is S (n), non-required parameter signal is N (n). then integrated signal is

X (n)=S (n)+N (n) n=0,1 ... .., M-1；

∑ N is had for N (n)_i(n)=0i=0,1 ..., k, k → ∞.

Signal AX (n) after obtaining superposed average after carrying out abundant superposed average number of times is

AX (n)=(∑ S_i(n)+∑N_i(n))/k, i=0,1,2 ..., k,

K is positive integer, when k → ∞ has, and ∑ S_i(n)=kS_i(n), so AX (n)=S_i(n)

If N_iN () average is not 0, and be C (n), then ∑ N_i(n)/k=C (n), i=0,1,2.... ∞；C (n) can be constant, represents that interference signal has a DC component, now AX (n)=S_i(n)+C (n), i=0,1,2.... ∞.

Adopt said method, other signals except 8 required parameter signals can be filtered, reduce signal noise.

Beneficial effect: adopt method for building up and the system of the sclerosis of blood vessels characteristic spectrum multiparameter medical model of the present invention, step is simple, it is accurate to calculate, degree of accuracy is high and the advantage of low cost；The healthy population that can calculate according to this model and the model result of typical blood vessel wall Disease, set up the data base for the monitoring of population health degree, and data acquisition amount is big, and making data base is large database concept.

Accompanying drawing explanation

Fig. 1 is the method flow diagram of the present invention；

Fig. 2 is the system construction drawing of the present invention.

Detailed description of the invention

Below in conjunction with embodiment and accompanying drawing, the invention will be further described.

As it is shown in figure 1, the method for building up of a kind of sclerosis of blood vessels characteristic spectrum multiparameter medical model, comprise the following steps:

Step 1: transfer the dynamic implicit information of multiparameter inside and outside blood vessel wall from data base, the dynamic implicit information of described multiparameter includes the acceleration wave of blood vessel wall line pressure shift ripple, blood vessel wall line pressure shift wave velocity ripple, blood vessel wall pressure wave acceleration wave, cardiechema signals, electrocardio ripple, capacity of blood vessel ripple, the differential ripple of capacity of blood vessel ripple, capacity of blood vessel ripple.Wherein electrocardio ripple is the reference signal gathered, blood vessel wall line pressure shift ripple, blood vessel wall line pressure shift wave velocity ripple, blood vessel wall pressure wave acceleration wave are the external dynamic information of blood vessel wall, and cardiechema signals, capacity of blood vessel ripple, the differential ripple of capacity of blood vessel ripple, cardiechema signals are blood vessel wall internal dynamic information；

Step 2: adopt fast Fourier transform that the dynamic implicit information of multiparameter gathered carries out the conversion of time domain and frequency domain in conjunction with the method for temporal decimation.

Step 3: adopt following steps to be analyzed the seasonal effect in time series every kind dynamic state of parameters implicit information that has collected, obtain the general features of angiopathy:

Step is 1., to adopt the method for rolling average to carry out its mathematic(al) representation of trend analysis every kind of parameter as follows:

$\frac{t_1+t_2+\mathrm{...}+t_n}{n},\frac{t_2+t_3+\mathrm{...}+t_{n+1}}{n},\frac{t_3+t_4+\mathrm{...}+t_{n+2}}{n},\frac{t_4+t_5+\mathrm{...}+t_{n+3}}{n}$

Wherein t_nRepresenting the data value in the n-th moment, n represents the data bulk of collection.Eliminate the irregular variation in every kind of parameter, set up mathematical model of the computer；

Step 2., adopt below equation every kind of parameter is carried out correlation analysis, obtain the universal law of disease

$r_{A,B}=\frac{\underset{t=1}{\overset{T}{\Σ}}(a_t-\stackrel{\‾}{A})(b_t-\stackrel{\‾}{B})}{{\mathrm{T\σ}}_A{\mathrm{\σ}}_B}$

Wherein, r_A,BRepresent A, B two relative coefficient of track data；a_t,b_tRepresent the t moment A, B two data value of track data respectively；T represents total time series number；Represent the average of data A and data B respectively；σ_A,σ_BRepresent the standard variance of data A and data B respectively, if r_A,BEqual to 0, then data A and data B is uncorrelated, otherwise is then correlated with.

Step 3., adopt the classification in data warehouse and data mining technology, cluster, correlation rule, neutral net, genetic algorithm that implicit information data dynamic inside and outside blood vessel wall are further analyzed, set up data mining model.

Step 4: adopt period map method to excavate the data of every kind of parameter of the dynamic implicit information of multiparameter carry out power Spectral Estimation, set up the power Spectral Estimation p of the characteristic spectrum of every kind of parameter_PER(w), its formula is:

$P_{PER}\left(w\right)=\frac{1}{N}{\left|\underset{n=0}{\overset{n-1}{\Σ}}x\left(n\right)e^{-j\mathrm{\π}wn}\right|}^2=\frac{1}{N}{\left|X_N\left(w\right)\right|}^2$

X_NW () represents the frequency domain sequence of every kind of parameter；X (n) represents the frequency domain sequence of every kind of parameter, P_PERW () represents that the power of the characteristic spectrum of every kind of parameter is estimated, N represents the observation station quantity of every kind of input parameter signal.

Step 5: obtaining testing result H according to the power Spectral Estimation of every kind of parameter and sclerosis of blood vessels characteristic spectrum multiparameter medical model, the formula of this multiparameter medical model is:

$H=\underset{i=1}{\overset{8}{\Σ}}{k}_i{P}_{PER\left(w\right)}$

Wherein, k_iRepresent the weighter factor of every kind of parameter, P_PERW () represents the power Spectral Estimation of every kind of parameter；The weighter factor k of every kind of parameter_iAccording to following table value.

Parameter type	Weighter factor kiValue
		Electrocardio ripple	2.0
Blood vessel wall line pressure shift ripple	3.0
		Blood vessel wall line pressure shift wave velocity ripple	2.5
Blood vessel wall pressure wave acceleration wave	2.5
		Cardiechema signals	1.5
Capacity of blood vessel ripple	3.0
		The differential ripple of capacity of blood vessel ripple	2.5
The second-order differential ripple of capacity of blood vessel ripple	2.0

According to above-mentioned steps, transfer the dynamic implicit information of blood vessel wall of M healthy individuals, and obtain healthy population vessel walls degree average horizontal value H according to formula (1)₀, then according to above-mentioned steps draws the vessel walls degree value of calculation H of every kind of typical blood vessel wall diseased individuals_i, according to formula (2), draw the vessel walls degree fiducial value P of every kind of typical blood vessel wall diseased individuals_i, formula expression is:

$H_0=\frac{\underset{m=1}{\overset{M}{\Σ}}{H}_m}{M}---\left(1\right)$

$P_i=\frac{H_0-H_i}{H_0}\×100\%---\left(2\right)$

According to the method described above, every kind of typical blood vessel wall disease gathers X the individual dynamic implicit information of blood vessel wall, and obtains the vessel walls degree fiducial value P of each individuality_i, then according to all of vessel walls degree fiducial value P_iSetting up a large database concept, this data base is for the big data monitoring for population health degree.

As shown in Figure 2, a kind of software system based on sclerosis of blood vessels characteristic spectrum multiparameter medical model, including data acquisition module, individual management module, signal processing module, system management module, data memory module, the data of data collecting module collected include individual data and the dynamic implicit information signal of multiparameter, and individual data and the dynamic implicit information signal of multiparameter are sent respectively to case management module and signal processing module, individual management module sets up individual archives, the dynamic implicit information signal of multiparameter is processed by signal processing module, data after process are sent respectively to case management module and data memory module, and it is sent to data memory module storage, system management module management data collection module, case management module, signal processing module and data memory module.

Described case management module includes case enquiry module, data editor module, transfer the files analysis module and report print module；Described signal processing module includes signal filtering module and signal analysis module, it is shown that signal filtering module adopts following methods that signal is filtered:

Assuming that desired parameters signal is S (n), non-required parameter signal is N (n). then integrated signal is

X (n)=S (n)+N (n) n=0,1 ... .., M-1；

∑ N is had for N (n)_i(n)=0i=0,1 ..., k, k → ∞.

Signal AX (n) after obtaining superposed average after carrying out abundant superposed average number of times is

AX (n)=(∑ S_i(n)+∑N_i(n))/k, i=0,1,2 ..., k,

K is positive integer, when k → ∞ has, and ∑ S_i(n)=kS_i(n), so AX (n)=S_i(n)

If N_iN () average is not 0, and be C (n), then ∑ N_i(n)/k=C (n), i=0,1,2.... ∞；C (n) can be constant, represents that interference signal has a DC component, now AX (n)=S_i(n)+C (n), i=0,1,2.... ∞.

Described signal analysis module adopts the medical model that the method for building up of described sclerosis of blood vessels characteristic spectrum multiparameter medical model is set up to be analyzed；Described system management module includes that display arranges module, event arranges module, system setup module and waveform display module；Described memory module includes signal memory module and case memory module.

Finally it should be noted that; foregoing description is only the preferred embodiments of the present invention; those of ordinary skill in the art is under the enlightenment of the present invention; under the premise without prejudice to present inventive concept and claim; can making and representing like multiple types, such conversion each falls within protection scope of the present invention.

Claims (8)

1. the method for building up of a sclerosis of blood vessels characteristic spectrum multiparameter medical model, it is characterised in that: method for building up comprises the following steps:

Step 1: transfer the dynamic implicit information of multiparameter inside and outside blood vessel wall from data base, the dynamic implicit information of described multiparameter includes the acceleration wave of blood vessel wall line pressure shift ripple, blood vessel wall line pressure shift wave velocity ripple, blood vessel wall pressure wave acceleration wave, cardiechema signals, electrocardio ripple, capacity of blood vessel ripple, the differential ripple of capacity of blood vessel ripple, capacity of blood vessel ripple；

Wherein electrocardio ripple is the reference signal gathered, blood vessel wall line pressure shift ripple, blood vessel wall line pressure shift wave velocity ripple, blood vessel wall pressure wave acceleration wave are the external dynamic information of blood vessel wall, and cardiechema signals, capacity of blood vessel ripple, the differential ripple of capacity of blood vessel ripple, cardiechema signals are blood vessel wall internal dynamic information；

Step 2: the signal of dynamic for the multiparameter transferred implicit information is transformed from the time domain to frequency domain；

Step 3: every kind of dynamic state of parameters implicit information through frequency domain transform is analyzed；

Step 4: the characteristic spectrum of every kind of parameter is carried out power Spectral Estimation, sets up the power Spectral Estimation p of the characteristic spectrum of every kind of parameter_PER(w)；

Step 5: obtaining testing result according to the power Spectral Estimation of every kind of parameter and sclerosis of blood vessels characteristic spectrum multiparameter medical model, the formula of this multiparameter medical model is:

$H=\underset{i=1}{\overset{8}{\Σ}}{k}_i{P}_{PER\left(w\right)}$

Wherein, k_iRepresent the weighter factor of every kind of parameter, p_PER(w)Represent the power Spectral Estimation of every kind of parameter.

2. the method for building up of sclerosis of blood vessels characteristic spectrum multiparameter medical model according to claim 1, it is characterised in that: step 2 adopt fast Fourier transform the dynamic implicit information of multiparameter gathered carries out the conversion of time domain and frequency domain in conjunction with the method for temporal decimation.

3. the method for building up of sclerosis of blood vessels characteristic spectrum multiparameter medical model according to claim 1, it is characterised in that: step 3 adopts following steps each parameter is analyzed:

Step 1., every kind of parameter is carried out trend analysis, eliminates the irregular variation in every kind of parameter, sets up mathematical model of the computer；

Step 2., adopt below equation every kind of parameter is carried out correlation analysis, obtain the dependency between every kind of parameter:

$r_{A,B}=\frac{\underset{t=1}{\overset{T}{\Σ}}(a_t-\stackrel{\‾}{A})(b_t-\stackrel{\‾}{B})}{{\mathrm{T\σ}}_A{\mathrm{\σ}}_B}$

Wherein, r_A,BRepresent the relative coefficient of two kinds of supplemental characteristics of A, B；a_t,b_tRepresent the data value of two kinds of supplemental characteristics of the t moment A, B respectively；T represents total time series number；Represent the average of supplemental characteristic A and supplemental characteristic B respectively；σ_A,σ_BRepresent the standard variance of supplemental characteristic A and supplemental characteristic B respectively, if r_A,BEqual to 0, then supplemental characteristic A and supplemental characteristic B is uncorrelated, otherwise is then correlated with.

4. the method for building up of sclerosis of blood vessels characteristic spectrum multiparameter medical model and software system thereof according to claim 1, it is characterized in that: adopting period map method that the characteristic spectrum of every kind of parameter of every kind of dynamic state of parameters implicit information is carried out power Spectral Estimation in step 4, its formula is:

$P_{PER}\left(w\right)=\frac{1}{N}|\underset{n=0}{\overset{n-1}{\mathrm{\Σ}}}x\left(n\right)e^{-j\mathrm{\π}wn}{|}^2=\frac{1}{N}|X_N\left(w\right){|}^2$

X_NW () represents the frequency domain sequence of every kind of parameter；X (n) represents the time domain sequences of every kind of parameter, P_PERW () represents that the power of the characteristic spectrum of every kind of parameter is estimated, N represents the observation station quantity of every kind of input parameter signal.

5. the method for building up of sclerosis of blood vessels characteristic spectrum multiparameter medical model according to claim 3, it is characterised in that: it is as follows that 1. step adopts the method for rolling average to carry out its mathematic(al) representation of trend analysis:

$\frac{t_1+t_2+...+t_n}{n},\frac{t_2+t_3+...+t_{n+1}}{n},\frac{t_3+t_4+...+t_{n+2}}{n},\frac{t_4+t_5+...+t_{n+3}}{n}$

Wherein t_nRepresenting the data value in the n-th moment, n represents the data bulk of collection.

6. the method for building up of sclerosis of blood vessels characteristic spectrum multiparameter medical model according to claim 1, it is characterised in that: step 3 adopt the classification in data warehouse and data mining technology, cluster, correlation rule, neutral net, genetic algorithm implicit information data dynamic inside and outside blood vessel wall are further analyzed.

7. the software system based on sclerosis of blood vessels characteristic spectrum multiparameter medical model, it is characterized in that: include data acquisition module, case management module, signal processing module, system management module and data memory module, the dynamic implicit information of blood vessel wall multiparameter of patient and patient data in described data collecting module collected hospital database, and deliver the information to signal processing module and be analyzed, data acquisition module and signal processing module send individual management module to personal information and analysis result respectively and form individual data, data memory module stores with individual data gathering data, system is configured and manages by system management module；

Described signal processing module is made up of signal filtering module and signal analysis module, the signal gathered is filtered by described signal filtering module, and described signal analysis module utilizes hardening characteristics spectrum multiparameter medical model that the multiparameter multidate information after after filtering is analyzed.

8. according to claim 7 based on the software system of sclerosis of blood vessels characteristic spectrum multiparameter medical model, it is characterised in that: described signal filtering module adopts following methods that signal is filtered:

Assuming that desired parameters signal is S (n), non-required parameter signal is N (n). then integrated signal is

X (n)=S (n)+N (n) n=0,1 ... .., M-1；

Σ N is had for N (n)_i(n)=0i=0,1 ..., k, k → ∞.

Signal AX (n) after obtaining superposed average after carrying out abundant superposed average number of times is

AX (n)=(Σ S_i(n)+ΣN_i(n))/k, i=0,1,2 ..., k,

K is positive integer, when k → ∞ has, and Σ S_i(n)=kS_i(n), so AX (n)=S_i(n)

If N_iN () average is not 0, and be C (n), then Σ N_i(n)/k=C (n), i=0,1,2.... ∞；C (n) can be constant, represents that interference signal has a DC component, now AX (n)=S_i(n)+C (n), i=0,1,2.... ∞.