CN103505248A - Ultrasonic work platform for monitoring renal blood flow through gastric ultrasound - Google Patents
Ultrasonic work platform for monitoring renal blood flow through gastric ultrasound Download PDFInfo
- Publication number
- CN103505248A CN103505248A CN201310431403.7A CN201310431403A CN103505248A CN 103505248 A CN103505248 A CN 103505248A CN 201310431403 A CN201310431403 A CN 201310431403A CN 103505248 A CN103505248 A CN 103505248A
- Authority
- CN
- China
- Prior art keywords
- data
- blood flow
- ultrasound
- ultrasonic
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention relates to the field of medical monitoring, in particular to an ultrasonic work platform for monitoring renal blood flow through gastric ultrasound. The work platform comprises a computer and color Doppler ultrasound equipment, wherein the computer is connected with the color Doppler ultrasound equipment through a data line. The computer comprises an ultrasonic data transmission module, an ultrasound data analyzing module, a data processing module and a data display module, wherein the ultrasonic data transmission module receives signals of the color Doppler ultrasound equipment and is in signal connection with the ultrasound data analyzing module, the ultrasound data analyzing module is in signal connection with the data processing module, and the data processing module is in signal connection with the data display module. The worktable can extract ultrasonic image signals from an ultrasonic instrument, identify the signals, convert the signals into numerical signals and waveform images, output the numerical signals and the waveform images and display the numerical signals and the waveform images on a displayer in a moving waveform and real-time numerical values.
Description
Technical field
The present invention relates to medical monitoring field, is a kind of ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow specifically.
Background technology
Acute renal failure is the common complication after extracorporeal circulation operation on vessels of heart, its incidence rate approximately 1% ~ 30%, simultaneously, the interference to body due to anesthesia and operation, will cause hemodynamic serious disorder, if clinical monitoring or deal with improperly, can cause serious consequence, after renal failure, mortality rate reaches as high as 80%.At least 50% acute renal failure can directly damage owing to renal ischemic.Owing to lacking so far, can in time, accurately monitor the effective means of operation middle kidney blood flow, guarantee that in operation, the sufficient blood perfusion of kidney still faces a severe challenge.Therefore, seek a kind of work platforms that can effectively monitor operation middle kidney blood flow and there is important clinical meaning.
In prior art, also there is pair improvement that blood flow monitoring work platforms carries out, as number of patent application, be CN87212236, the applying date to be that 1987.8.24, name are called the Chinese utility model patent of " Computerized automatic monitor of blood flow dynamics ", its technical scheme is as follows: a kind of human blood flowing dynamics computer automatic mornitoring instrument.This instrument is comprised of computer system control circuit and measuring system, computer system is connected with measuring system by control circuit, computer system starts corresponding application program according to the predefined guardianship content of user, can carry out the automatic real-time measurement of human heart rate and human blood-pressure and overload alarm; Automatically the real-time and exceptional value prompting of the hemodynamic parameter of the organs such as Dui Ren body-centered, brain, lung, liver, kidney.By the exportable diagnosis report of printer being connected with computer, also patient's diagnosis report deposit can be filed.
Above-mentioned patent is to utilize electrode, admittance instrument and air pump to obtain the mobile signal of telecommunication of painstaking effort, can not realize in time, accurately monitor the object of operation middle kidney blood flow.
Summary of the invention
The problems referred to above that exist in order to overcome existing watch-dog, a kind of ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow of special proposition now.
For realizing above-mentioned technique effect, concrete scheme of the present invention is as follows:
A ultrasonic work platforms for stomach-ultrasonography monitoring kidney blood flow, is characterized in that: comprise computer and color ultrasound equipment, described computer is connected with color ultrasound equipment by data wire;
Described computer comprises ultrasound data transport module, ultrasound data parsing module, data processing module and data disaply moudle, described ultrasound data transport module receives the signal of color ultrasound equipment, and be connected with ultrasound data parsing module signal, described ultrasound data parsing module is connected with data processing module signal, and described data processing module is connected with data disaply moudle signal.
Between the described interface that utilizes color ultrasound equipment and described ultrasound data transport module, by data wire, be connected, the kidney blood flow data collecting is transferred to the ultrasound data parsing module on the computer that this platform moves, and kidney blood flow data comprises the image information of maximal rate, average speed, minimum speed and the renal artery impedance index as calculated of renal artery internal diameter, artery blood flow.
Described image information is decoded as numerical information by linear classification recognition methods, comprise PCA: step, for the multidimensional variable of the image information of the maximal rate that comprises renal artery internal diameter, artery blood flow, average speed, minimum speed and renal artery impedance index is as calculated carried out to dimension-reduction treatment, makes above multidimensional variable convert low-dimensional variable under the prerequisite that keeps precision.
Described image information is decoded as numerical information by linear classification recognition methods, comprise linear discriminant analysis: step is in the inseparable data projection to of the multidimensional variable neutral line direction of image information of the maximal rate, average speed, minimum speed and the renal artery impedance index as calculated that comprise renal artery internal diameter, artery blood flow, make the data after projection be similar to linear separability, data difference between inhomogeneity target individual after projection increases, and reduces the difference between similar target individual simultaneously.
Described image information is decoded as numerical information by linear classification recognition methods, comprise principal component analysis+linear discriminant analysis: step, for first random vector being carried out to principal component analysis to reduce its dimension, is then carried out linear discriminant analysis to the later random vector of dimensionality reduction.
Described ultrasound data parsing module, by reading the data of ultrasound data transport module transmission, parses renal artery internal diameter, kidney blood flow maximal rate, average speed, minimum speed, Velocity-time integration and hrv parameter.
This platform, by image identification function, ultrasonic pictorial information, resolves to numerical information, specifically comprises:
To what collect, by the measured initial data of ultrasound wave, carry out Classification and Identification, and extract useful information (initial data being carried out to pretreatment), adopt following three kinds of classifying identification methods:
(1) linear classification recognition methods;
(2) Nonlinear Classification recognition methods;
(3) method of least square support vector machine (LS-SVM).
Described linear classification recognition methods comprises (a) principal component analysis (PCA, principal component analysis), linear discriminant analysis (LDA, Linear Discriminant Analysis) claim again Fisher linear discriminant (FLD), (c) PCA+LDA algorithm.
Described data disaply moudle is made oscillogram by individual parameter, shows in real time.
The invention has the advantages that:
1, work platforms of the present invention can extract ultrasound image signal from ultrasonic instrument, and identifies, then converts numerical signal and mode chart picture to, and then output shows with mobile wave mode and real-time numerical value on display.
2, by the present invention, can see real-time kidney blood circumstance in art, thus assessment renal perfusion situation.In operation, do not monitor this project of kidney blood flow at present.Kidney blood flow can be monitored through transabdominal ultrasonography, but can disturb operation technique, so also cannot Real-Time Monitoring kidney blood flow in operation at present.The object of monitoring kidney blood flow is because kidney perfusion deficiency is the major reason that causes postoperative death and renal dysfunction.If can find in time that kidney perfusion is not enough, can prevention of postoperative renal dysfunction.
3, compare with existing similar work platforms, what existing platform was used is electrode, admittance instrument and air pump; And it is used in this application ultrasonic.The signal that existing platform obtains is the signal of telecommunication; What the application obtained is ultrasonic signal.Existing platform object is integrated circuit, monitoring electrocardio, blood pressure and each organ admittance; The application's object is monitoring kidney blood flow parameter (comprising flow velocity, flow and kidney vascular hindrance index).
4,, by principal component analysis, can, so that data are processed in low dimensional feature space, keep again most information in initial data simultaneously.By LDA, convert, can further reduce the dimension of luv space data characteristics, can make dispersion maximum between the class of data after projection simultaneously, in class, dispersion is minimum, thereby improves classifying quality.
Accompanying drawing explanation
Fig. 1 is structural representation of the present invention.
The specific embodiment
Ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow comprises computer and color ultrasound equipment, and described computer is connected with color ultrasound equipment by data wire; Described computer comprises ultrasound data transport module, ultrasound data parsing module, data processing module and data disaply moudle, described ultrasound data transport module receives the signal of color ultrasound equipment, and be connected with ultrasound data parsing module signal, described ultrasound data parsing module is connected with data processing module signal, and described data processing module is connected with data disaply moudle signal.Between the described interface that utilizes color ultrasound equipment and described ultrasound data transport module, by data wire, be connected, the kidney blood flow data collecting is transferred to the ultrasound data parsing module on the computer that this platform moves, and kidney blood flow data comprises the image information of maximal rate, average speed, minimum speed and the renal artery impedance index as calculated of renal artery internal diameter, artery blood flow.
Image information is decoded as numerical information by linear classification recognition methods, comprise PCA: step, for the multidimensional variable of the image information of the maximal rate that comprises renal artery internal diameter, artery blood flow, average speed, minimum speed and renal artery impedance index is as calculated carried out to dimension-reduction treatment, makes above multidimensional variable convert low-dimensional variable under the prerequisite that keeps precision.Described image information is decoded as numerical information by linear classification recognition methods, comprise linear discriminant analysis: step is in the inseparable data projection to of the multidimensional variable neutral line direction of image information of the maximal rate, average speed, minimum speed and the renal artery impedance index as calculated that comprise renal artery internal diameter, artery blood flow, make the data after projection be similar to linear separability, data difference between inhomogeneity target individual after projection increases, and reduces the difference between similar target individual simultaneously.
Image information is decoded as numerical information by linear classification recognition methods, comprises principal component analysis+linear discriminant analysis: step, for first random vector being carried out to principal component analysis to reduce its dimension, is then carried out linear discriminant analysis to the later random vector of dimensionality reduction.Described ultrasound data parsing module, by reading the data of ultrasound data transport module transmission, parses renal artery internal diameter, kidney blood flow maximal rate, average speed, minimum speed, Velocity-time integration and hrv parameter.This platform, by image identification function, ultrasonic pictorial information, resolves to numerical information, specifically comprises:
To what collect, by the measured initial data of ultrasound wave, carry out Classification and Identification, and extract useful information (initial data being carried out to pretreatment), adopt following three kinds of classifying identification methods:
(1) linear classification recognition methods;
(2) Nonlinear Classification recognition methods;
(3) method of least square support vector machine (LS-SVM).
Described linear classification recognition methods comprises (a) principal component analysis (PCA, principal component analysis), linear discriminant analysis (LDA, Linear Discriminant Analysis) claim again Fisher linear discriminant (FLD), (c) PCA+LDA algorithm.Described data disaply moudle is made oscillogram by individual parameter, shows in real time.
(a) principal component analysis (PCA, principal component analysis).It is a kind of common dimensionality reduction statistical method.It is by means of an orthogonal transformation, and the former random vector that its component is relevant converts the incoherent new random vector of its component to.This shows as and converts the covariance matrix of former random vector to diagonal matrix on algebraically; Convert former coordinate system to new orthogonal coordinate system geometrically showing as, make it to point to sample point and scatter several orthogonal directions of opening most, be that random vector carries out projection on these orthogonal directions, first variance is on first coordinate axes (being called first principal component or the first pivot direction), second largest variance is upper at second coordinate axes (Second principal component, or the second pivot direction), the like.Then by certain strategy, multidimensional variable system is carried out to suitable dimension-reduction treatment, make it to convert low-dimensional variable system to a higher precision.By principal component analysis, can, so that data are processed in low dimensional feature space, keep again most information in initial data simultaneously.
(b) linear discriminant analysis (LDA, Linear Discriminant Analysis) claim again Fisher linear discriminant (FLD).It is a kind of good feature extracting method.The basic thought of LDA conversion is by the inseparable data projection to of a higher dimensional space neutral line direction, make the data after projection be similar to linear separability, and the data after projection can increase the difference between inhomogeneity target individual, reduce the difference between similar target individual simultaneously.By LDA, convert, can further reduce the dimension of luv space data characteristics, can make dispersion maximum between the class of data after projection simultaneously, in class, dispersion is minimum, thereby improves classifying quality.
(c) PCA+LDA algorithm.First random vector is carried out to principal component analysis to reduce its dimension, then the later random vector of dimensionality reduction is carried out to linear discriminant analysis.The benefit of doing is like this, the advantages of PCA and LDA algorithm can be got up, to improve Classification and Identification effect.
Linear classification recognition methods is simple, easily operation, and algorithm real-time is good, but not high enough in some occasion classification and recognition.
Nonlinear Classification recognition methods:
The thought of core (Kernel) function is incorporated into respectively in PCA and LDA algorithm, thereby forms core principle component analysis (KPCA) and kernel Fisher discriminant analysis (KFDA) algorithm.KPCA and KFDA algorithm belong to Nonlinear Classification recognition methods, and its advantage is that algorithm performance is stable, and classification and recognition is high, but algorithm real-time is not as linear classification recognition methods.
Method of least square support vector machine (LS-SVM):
The core concept of least square method supporting vector machine is exactly to have replaced the inequality constraints in standard support vector machine by equality constraint.Owing to having adopted equality constraint, so originally need to solve the optimization problem of a quadratic programming, just converted to and solved a system of linear equations, like this, solve difficulty and greatly reduce.LS-SVM algorithm performance is stable, classification and recognition is not less than standard support vector machine (in a lot of occasions, the classification and recognition of LS-SVM algorithm will be higher than standard support vector machine), and algorithm operating is easily more a lot of than standard support vector machine, but algorithm real-time is not as linear classification recognition methods.
Above-mentioned classifying identification method is very classical mode identification method, by a large amount of facts have proved, is had very high feasibility.Specifically, in operation, can need to choose suitable appropriate classifying identification method according to different, thereby reach, initial data be carried out to accurate, real-time pretreated object.
Claims (8)
1. through a ultrasonic work platforms for stomach-ultrasonography monitoring kidney blood flow, it is characterized in that: comprise computer and color ultrasound equipment, described computer is connected with color ultrasound equipment by data wire;
Described computer comprises ultrasound data transport module, ultrasound data parsing module, data processing module and data disaply moudle, described ultrasound data transport module receives the signal of color ultrasound equipment, and be connected with ultrasound data parsing module signal, described ultrasound data parsing module is connected with data processing module signal, and described data processing module is connected with data disaply moudle signal;
Between the described interface that utilizes color ultrasound equipment and described ultrasound data transport module, by data wire, be connected, the kidney blood flow data collecting is transferred to the ultrasound data parsing module on the computer that this platform moves, and kidney blood flow data comprises the image information of maximal rate, average speed, minimum speed and the renal artery impedance index as calculated of renal artery internal diameter, artery blood flow.
2. the ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow according to claim 1, it is characterized in that: described image information is decoded as numerical information by linear classification recognition methods, comprise PCA: step, for the multidimensional variable of the image information of the maximal rate that comprises renal artery internal diameter, artery blood flow, average speed, minimum speed and renal artery impedance index is as calculated carried out to dimension-reduction treatment, makes above multidimensional variable convert low-dimensional variable under the prerequisite that keeps precision.
3. the ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow according to claim 2, it is characterized in that: described image information is decoded as numerical information by linear classification recognition methods, comprise linear discriminant analysis: step is for comprising renal artery internal diameter, the maximal rate of artery blood flow, average speed, in the inseparable data projection to of a multidimensional variable neutral line direction of the image information of minimum speed and renal artery impedance index as calculated, make the data after projection be similar to linear separability, data difference between inhomogeneity target individual after projection increases, reduce the difference between similar target individual simultaneously.
4. the ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow according to claim 3, it is characterized in that: described image information is decoded as numerical information by linear classification recognition methods, comprise principal component analysis+linear discriminant analysis: step, for first random vector being carried out to principal component analysis to reduce its dimension, is then carried out linear discriminant analysis to the later random vector of dimensionality reduction.
5. the ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow according to claim 4, it is characterized in that: described ultrasound data parsing module, by reading the data of ultrasound data transport module transmission, parses renal artery internal diameter, kidney blood flow maximal rate, average speed, minimum speed, Velocity-time integration and hrv parameter.
6. the ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow according to claim 5, is characterized in that: this platform, by image identification function, ultrasonic pictorial information, resolves to numerical information, specifically comprises:
To what collect, by the measured initial data of ultrasound wave, carry out Classification and Identification, and extract useful information (initial data being carried out to pretreatment), adopt following three kinds of classifying identification methods:
(1 linear classification recognition methods;
(2 Nonlinear Classification recognition methodss;
(3 method of least square support vector machine.
7. the ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow according to claim 6, it is characterized in that: described linear classification recognition methods comprises (a) principal component analysis, linear discriminant analysis claims again Fisher linear discriminant (FLD), (c) PCA+LDA algorithm.
8. the ultrasonic work platforms through stomach-ultrasonography monitoring kidney blood flow according to claim 7, is characterized in that: described data disaply moudle is made oscillogram by individual parameter, shows in real time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310431403.7A CN103505248A (en) | 2013-09-22 | 2013-09-22 | Ultrasonic work platform for monitoring renal blood flow through gastric ultrasound |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310431403.7A CN103505248A (en) | 2013-09-22 | 2013-09-22 | Ultrasonic work platform for monitoring renal blood flow through gastric ultrasound |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103505248A true CN103505248A (en) | 2014-01-15 |
Family
ID=49889049
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310431403.7A Pending CN103505248A (en) | 2013-09-22 | 2013-09-22 | Ultrasonic work platform for monitoring renal blood flow through gastric ultrasound |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103505248A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6514208B1 (en) * | 1999-10-18 | 2003-02-04 | The United States Of America As Represented By The Secretary Of The Army | Method and apparatus for power doppler ultrasound image analysis |
| CN101187977A (en) * | 2007-12-18 | 2008-05-28 | 北京中星微电子有限公司 | A face authentication method and device |
| CN101210966A (en) * | 2006-12-30 | 2008-07-02 | 深圳迈瑞生物医疗电子股份有限公司 | Priority coding after-treatment method and device colorful blood stream imaging |
| CN103099642A (en) * | 2013-02-04 | 2013-05-15 | 左旺孟 | Ultrasonic blood flow signal quality real-time analysis method |
-
2013
- 2013-09-22 CN CN201310431403.7A patent/CN103505248A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6514208B1 (en) * | 1999-10-18 | 2003-02-04 | The United States Of America As Represented By The Secretary Of The Army | Method and apparatus for power doppler ultrasound image analysis |
| CN101210966A (en) * | 2006-12-30 | 2008-07-02 | 深圳迈瑞生物医疗电子股份有限公司 | Priority coding after-treatment method and device colorful blood stream imaging |
| CN101187977A (en) * | 2007-12-18 | 2008-05-28 | 北京中星微电子有限公司 | A face authentication method and device |
| CN103099642A (en) * | 2013-02-04 | 2013-05-15 | 左旺孟 | Ultrasonic blood flow signal quality real-time analysis method |
Non-Patent Citations (1)
| Title |
|---|
| 杨平亮等: "经食管超声监测心血管术中肾血流的应用研究", 《四川大学学报》, vol. 40, no. 1, 15 January 2009 (2009-01-15), pages 166 - 169 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109886933B (en) | Medical image recognition method and device and storage medium | |
| US11229419B2 (en) | Method for processing 3D image data and 3D ultrasonic imaging method and system | |
| Gudi et al. | Real-time webcam heart-rate and variability estimation with clean ground truth for evaluation | |
| CN108364290B (en) | Method, medium, and system for analyzing a sequence of images of periodic physiological activity | |
| CN103654760B (en) | Noninvasive intracranial pressure measuring method and the Noninvasive intracranial pressure analyser of application the method | |
| KR102676569B1 (en) | Medical image analysis apparatus and method, medical image visualization apparatus and method | |
| CN105764412B (en) | The natural frequencies analysis determined for left ventricular ejection fraction or stroke volume | |
| CN107169419A (en) | Contactless humanbody sign detection method and device based on machine vision | |
| CN112597982B (en) | Image classification method, device, equipment and medium based on artificial intelligence | |
| Premkumar et al. | Intelligent remote photoplethysmography-based methods for heart rate estimation from face videos: A survey | |
| EP4193913A1 (en) | Vital sign detection method and electronic device | |
| CN110246135A (en) | Monitor Follicles method, apparatus, system and storage medium | |
| CN202950650U (en) | Non-invasive intracranial pressure analysis meter and non-invasive intracranial pressure detection system | |
| CN113693611A (en) | Machine learning-based electrocardiogram data classification method and device | |
| Ontiveros et al. | Evaluating RGB channels in remote photoplethysmography: a comparative study with contact-based PPG | |
| CN103505248A (en) | Ultrasonic work platform for monitoring renal blood flow through gastric ultrasound | |
| KR102488569B1 (en) | A method for determining a object target of the body that minimizes the risk of target deviation, a method for calculating the center coordinates of a final object target, an automatic blood collection device using the same, and a program stored in a computer-readable recording medium | |
| US20220249060A1 (en) | Method for processing 3d image data and 3d ultrasonic imaging method and system | |
| CN110648328B (en) | Aorta replacement surgery remote intelligent screening system based on 5G | |
| CN112819786B (en) | Liver hemodynamic detection device based on multiple hepatic vein wave patterns | |
| CN114758400A (en) | Non-contact human body blood oxygen saturation detection method, device and equipment | |
| CN108498088A (en) | A kind of instrument of detection and analysis cerebral artery function and state | |
| CN114092425A (en) | Cerebral ischemia scoring method based on diffusion weighted image, electronic device and medium | |
| CN109171807B (en) | Signal processing method and system of ultrasonic diagnostic equipment and ultrasonic diagnostic equipment | |
| CN109009020A (en) | A kind of health monitoring systems and health monitor method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140115 |