CN113724208A - Blood flow frequency spectrum signal classification method and system - Google Patents
Blood flow frequency spectrum signal classification method and system Download PDFInfo
- Publication number
- CN113724208A CN113724208A CN202110928460.0A CN202110928460A CN113724208A CN 113724208 A CN113724208 A CN 113724208A CN 202110928460 A CN202110928460 A CN 202110928460A CN 113724208 A CN113724208 A CN 113724208A
- Authority
- CN
- China
- Prior art keywords
- blood flow
- spectrum
- transfer function
- image
- classification
- 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.)
- Granted
Links
- 238000001228 spectrum Methods 0.000 title claims abstract description 105
- 230000017531 blood circulation Effects 0.000 title claims abstract description 95
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000012546 transfer Methods 0.000 claims abstract description 60
- 238000012545 processing Methods 0.000 claims abstract description 48
- 230000008321 arterial blood flow Effects 0.000 claims abstract description 14
- 238000004458 analytical method Methods 0.000 claims abstract description 11
- 239000008280 blood Substances 0.000 claims abstract description 11
- 210000004369 blood Anatomy 0.000 claims abstract description 11
- 238000000605 extraction Methods 0.000 claims abstract description 10
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 8
- 238000004364 calculation method Methods 0.000 claims abstract description 4
- 210000002302 brachial artery Anatomy 0.000 claims description 30
- 210000001715 carotid artery Anatomy 0.000 claims description 20
- 238000011049 filling Methods 0.000 claims description 12
- 238000003708 edge detection Methods 0.000 claims description 10
- 238000002604 ultrasonography Methods 0.000 claims description 7
- 210000001367 artery Anatomy 0.000 claims description 6
- 230000008030 elimination Effects 0.000 claims description 6
- 238000003379 elimination reaction Methods 0.000 claims description 6
- 239000011800 void material Substances 0.000 claims description 5
- 230000003595 spectral effect Effects 0.000 claims description 3
- 230000002194 synthesizing effect Effects 0.000 claims description 3
- 238000010276 construction Methods 0.000 abstract 1
- 239000000284 extract Substances 0.000 description 5
- 210000001168 carotid artery common Anatomy 0.000 description 4
- 238000001595 flow curve Methods 0.000 description 4
- 208000029078 coronary artery disease Diseases 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000005284 excitation Effects 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000008081 blood perfusion Effects 0.000 description 1
- 208000026106 cerebrovascular disease Diseases 0.000 description 1
- 230000004087 circulation Effects 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000000004 hemodynamic effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 210000003739 neck Anatomy 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000002560 therapeutic procedure Methods 0.000 description 1
- 210000001835 viscera Anatomy 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/06—Measuring blood flow
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B8/00—Diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/52—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves
- A61B8/5215—Devices using data or image processing specially adapted for diagnosis using ultrasonic, sonic or infrasonic waves involving processing of medical diagnostic data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10132—Ultrasound image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
- G06T2207/30104—Vascular flow; Blood flow; Perfusion
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Pathology (AREA)
- Public Health (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Hematology (AREA)
- Quality & Reliability (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention discloses a blood flow frequency spectrum signal classification method and a system, wherein the method comprises the following steps: s1, acquiring an arterial blood vessel ultrasonic image and extracting an arterial blood flow frequency spectrum change curve; s2, establishing an ARX transfer function of a blood velocity spectrum according to the arterial blood flow spectrum change curve; s3, performing fitting analysis on the ARX transfer function of the blood flow velocity spectrum to obtain the transfer function characteristics; and S4, processing the transfer function characteristics by adopting an SVM classifier to finish the classification of the blood flow spectrum signals. The system comprises: the device comprises a spectrum curve extraction module, a transfer function construction module, a feature calculation module and a classification module. By using the invention, accurate classification of signals is realized. The blood flow frequency spectrum signal classification method and system can be widely applied to the field of signal classification.
Description
Technical Field
The present invention relates to the field of signal classification, and in particular, to a method and a system for classifying blood flow spectrum signals.
Background
Ultrasound technology is widely used for medical imaging and measurement. External counterpulsation is one of noninvasive biofeedback technologies for effectively regulating and controlling the cardio-cerebral circulation hemodynamic environment, namely important visceral organ blood perfusion at present, and has great development potential in the field of cardio-cerebrovascular disease rehabilitation.
Disclosure of Invention
In order to solve the above technical problems, an object of the present invention is to provide a method and a system for classifying blood flow spectrum signals, which process dynamically monitored signals, so as to realize accurate classification of the signals.
The first technical scheme adopted by the invention is as follows: a blood flow spectrum signal classification method comprises the following steps:
s1, acquiring an arterial blood vessel ultrasonic image and extracting an arterial blood flow frequency spectrum change curve;
s2, establishing an ARX transfer function of a blood velocity spectrum according to the arterial blood flow spectrum change curve;
s3, performing fitting analysis on the ARX transfer function of the blood flow velocity spectrum to obtain the transfer function characteristics;
and S4, processing the transfer function characteristics by adopting an SVM classifier to finish the classification of the blood flow spectrum signals.
Further, the step of acquiring an arterial blood vessel ultrasound image and extracting an arterial blood flow frequency spectrum change curve specifically includes:
s11, acquiring a carotid artery blood vessel ultrasonic image and a brachial artery blood vessel ultrasonic image;
s12, carrying out interference factor elimination processing, binarization processing, void filling processing and Sobel operator edge detection processing on the carotid artery blood vessel ultrasonic image to obtain a first Sobel gradient image;
s13, extracting the first Sobel gradient image to obtain a carotid blood flow frequency spectrum change curve;
s14, performing region-of-interest selection, binarization processing, cavity filling processing and Sobel operator edge detection processing on the brachial artery blood vessel ultrasonic image to obtain a second Sobel gradient image;
and S15, identifying the 0-axis position in the second Sobel gradient image, and locking the K value to search the upper envelope and the lower envelope to obtain the brachial artery blood flow frequency spectrum change curve.
Further, the step of performing interference factor elimination processing, binarization processing, cavity filling processing and Sobel operator edge detection processing on the carotid artery blood vessel ultrasonic image to obtain a first Sobel gradient image specifically includes:
s121, setting the pixel value of the interference factor in the carotid artery blood vessel ultrasonic image as background black to obtain a carotid artery blood flow frequency spectrum image with the interference factor removed;
s122, converting the carotid artery blood flow spectrum image after the interference factors are removed into a binary image;
s123, carrying out hole filling processing on the binary image by adopting closed operation to obtain a filled image;
and S124, respectively calculating gradients of the filled images in the horizontal direction and the vertical direction by adopting a Sobel operator and synthesizing to obtain a first Sobel gradient image.
Further, the step of identifying the 0-axis position in the second Sobel gradient image, and locking the K value to search the upper envelope and the lower envelope to obtain the brachial artery blood flow frequency spectrum change curve specifically includes:
s151, searching pixel points with pixel values of 1 in the region of interest in the second Sobel gradient image in a global mode to obtain a position set of white pixel values;
s152, determining the position of the axis 0 in the region of interest according to the four highest longitudinal coordinate values in the position set;
s153, setting K to be 1, and extracting an upper envelope from top to bottom until a pixel point in the 0-axis range is identified;
s154, setting K to be-1, and extracting a lower envelope from bottom to top until a pixel point in a 0-axis range is identified;
and S155, returning to the step S153 until all pixel points in the region of interest are identified, and obtaining a brachial artery blood flow frequency spectrum change curve.
Further, the ARX transfer function of the blood velocity spectrum is specifically a 12-order ARX model, and the step of performing fitting analysis on the ARX transfer function of the blood velocity spectrum to obtain transfer function characteristics includes:
and fitting and analyzing the ARX transfer function of the blood flow velocity spectrum based on a least square method, and extracting coefficient characteristics of the transfer function, horizontal and vertical coordinate characteristics of a zero point and pole horizontal and vertical coordinate characteristics.
Further, the step of processing the transfer function features by using an SVM classifier to complete the classification of the blood flow spectrum signal specifically includes:
selecting the transfer function characteristics based on an LASSO regression algorithm to obtain selected pole horizontal and vertical coordinate characteristics;
classifying according to the selected zero-pole horizontal and vertical coordinate characteristics based on an SVM classifier to obtain a classification result;
and reflecting the corresponding blood flow spectrum signal classification according to the classification result.
The second technical scheme adopted by the invention is as follows: a blood flow spectral signal classification system comprising:
the spectrum curve extraction module is used for acquiring an artery blood vessel ultrasonic image and extracting an artery blood flow spectrum change curve;
the transfer function building module is used for building an ARX transfer function of a blood velocity spectrum according to the arterial blood flow spectrum change curve;
the characteristic calculation module is used for performing fitting analysis on an ARX transfer function based on the blood flow velocity spectrum to obtain transfer function characteristics;
and the classification module is used for processing the transfer function characteristics by adopting an SVM classifier to finish the classification of the blood flow spectrum signals.
The method and the system have the beneficial effects that: the invention extracts an accurate blood flow spectrum curve by preprocessing the blood vessel ultrasonic image, constructs an ARX model to calculate parameter characteristics, and finally realizes the classification of blood flow spectrum signals by classifying the parameter characteristics, and the classified signals can be used for assisting doctors to diagnose.
Drawings
FIG. 1 is a flow chart illustrating the steps of a blood flow spectrum signal classification method according to the present invention;
fig. 2 is a block diagram of a blood flow spectrum signal classification system according to the present invention.
Detailed Description
The invention is described in further detail below with reference to the figures and the specific embodiments. The step numbers in the following embodiments are provided only for convenience of illustration, the order between the steps is not limited at all, and the execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art.
Referring to fig. 1, the present invention provides a blood flow spectrum signal classification method, which includes the following steps:
s1, acquiring an arterial blood vessel ultrasonic image and extracting an arterial blood flow frequency spectrum change curve;
s2, establishing an ARX transfer function of a blood velocity spectrum according to the arterial blood flow spectrum change curve;
s3, performing fitting analysis on the ARX transfer function of the blood flow velocity spectrum to obtain the transfer function characteristics;
and S4, processing the transfer function characteristics by adopting an SVM classifier to finish the classification of the blood flow spectrum signals.
Further, as a preferred embodiment of the method, the step of acquiring an arterial blood vessel ultrasound image and extracting an arterial blood flow frequency spectrum change curve specifically includes:
s11, acquiring a carotid artery blood vessel ultrasonic image and a brachial artery blood vessel ultrasonic image;
specifically, ultrasound images of the brachial artery blood vessel of the sum of the front, middle and back necks of external counterpulsation (EECP) are acquired, and blood flow spectrum raw data in a DICOM format is derived.
S12, carrying out interference factor elimination processing, binarization processing, void filling processing and Sobel operator edge detection processing on the carotid artery blood vessel ultrasonic image to obtain a first Sobel gradient image;
s13, extracting the first Sobel gradient image to obtain a carotid blood flow frequency spectrum change curve;
s14, performing region-of-interest selection, binarization processing, cavity filling processing and Sobel operator edge detection processing on the brachial artery blood vessel ultrasonic image to obtain a second Sobel gradient image;
and S15, identifying the 0-axis position in the second Sobel gradient image, and locking the K value to search the upper envelope and the lower envelope to obtain the brachial artery blood flow frequency spectrum change curve.
Specifically, the blood flow velocity change pattern of the brachial artery and the common carotid artery are different, and the peripheral brachial artery blood vessel is mainly dominated by three-way blood flow.
Further, as a preferred embodiment of the method, the step of performing interference factor elimination processing, binarization processing, void filling processing and Sobel operator edge detection processing on the carotid artery blood vessel ultrasound image to obtain a first Sobel gradient image specifically includes:
the common carotid artery blood flow signal is relatively stable, and reverse blood flow does not exist in a normal state. The blood flow velocity frequency spectrum curve is distributed above the 0 axis, and the repeatability is good without processing such as filtering and the like.
S121, setting the pixel value of the interference factor in the carotid artery blood vessel ultrasonic image as background black to obtain a carotid artery blood flow frequency spectrum image with the interference factor removed;
specifically, for the reason of equipment, yellow star marks and cycle selection lines carried by the equipment are arranged on the acquired ultrasound image blood flow spectrum curve, and yellow lightning marks exist on a few images; these interference factors will become discontinuities on the envelope curve, causing large interference to the maximum velocity spectral curve extraction, thus affecting the data quality. According to the scheme, the yellow pixel point is set to be background black according to the pixel values (254,254 and 0) of the yellow pixel point, so that the threshold value of subsequent binarization processing is not influenced, and the interference of the yellow star mark and the selected line is eliminated.
S122, converting the carotid artery blood flow spectrum image after the interference factors are removed into a binary image;
specifically, in order to solve the problems that green outer envelope lines on yellow star marks in the image are discontinuous and key feature points are lacked, the common carotid artery blood flow spectrum image after interference factors are eliminated is converted into a binary image, and only black and white pixel points exist in the image. Therefore, the missing key feature points can be replaced by the white pixel points below, so that the time for finding points by manual difference is saved, and errors are obviously reduced.
S123, carrying out hole filling processing on the binary image by adopting closed operation to obtain a filled image;
specifically, after the closed operation setting, the blood flow frequency spectrum image edge is coherent and full, and preparation is made for extracting the outer edge in the next step. Through the closed operation, a large batch of images can be effectively extracted to the outer edge.
And S124, respectively calculating gradients of the filled images in the horizontal direction and the vertical direction by adopting a Sobel operator and synthesizing to obtain a first Sobel gradient image.
Specifically, the Sobel operator calculates the difference between the pixel points and the surrounding pixel points according to the gradient of each pixel point of the image, finally separates the target and the background, extracts the outer edge region, and includes two groups of 3 × 3 matrixes which are respectively horizontal and vertical, and performs plane convolution on the matrixes and the image to respectively obtain the brightness difference approximate values in the two directions.
In the above formula, A represents an original image, GxAnd GyRepresenting the gray values of the image detected by the transverse and longitudinal edges respectively
The horizontal and vertical gray values of each pixel of the image are combined by formula (2) to calculate the size of the dot gray:
if the gradient G is larger than a certain threshold value, the point (x, y) is considered as an edge point; the gradient direction can then be calculated using equation (3):
the Sobel operator is a relatively common effective edge detection method, detects an edge when the edge reaches an extreme value according to gray scale weighting difference of upper and lower and left and right adjacent points of a pixel point, calculates the difference between the pixel point and surrounding pixel points according to the gradient of each pixel point of an image, finally separates a target and a background, extracts the outer edge and obtains an accurate brachial artery blood flow maximum velocity curve.
Further, as a preferred embodiment of the method, the step of identifying the 0-axis position in the second Sobel gradient image, and locking the K value to search the upper envelope and the lower envelope to obtain the brachial artery blood flow spectrum variation curve specifically includes:
the blood flow velocity spectrum of the brachial artery is mostly three-way blood flow in the acquisition process, namely, the blood flow spectrum curve passes through the 0 axis, and the blood flow of the brachial artery is unstable before, during and after EECP therapy, and the scale is changeable. Since the brachial artery blood flow curve passes through the 0 axis, the maximum blood flow velocity curve should be composed of an upper envelope above the 0 axis and a lower envelope below the 0 axis. The maximum blood flow velocity curve extracted from the brachial artery should also include a line coincident with the 0 axis in the period, and therefore, the identification of the 0 axis position in the blood flow curve of the brachial artery is particularly important. However, the blood flow curve scale is variable, the distribution position of the blood flow spectrum curve in the image is not fixed, and the determination of the 0 axis becomes a key problem for extracting the blood flow curve of the brachial artery blood vessel.
S151, searching pixel points with pixel values of 1 in the region of interest in the second Sobel gradient image in a global mode to obtain a position set of white pixel values;
specifically, for brachial artery external envelope extraction algorithm, the study mainly sets a movable selection box. The selection box may drag a location to select a region of interest (ROI). However, the extraction algorithm needs to identify the 0-axis position first because of the difference of the blood flow direction. Since the pixel point of the 0 axis in the image is not a single straight line, but a wide area with a certain width on the ordinate. Therefore, when the position of the axis 0 is found, the range of the vertical coordinate of the axis 0 needs to be obtained, and the axis 0 is obtained through large-batch image observation and comprises the heights of four pixel points. In the algorithm for identifying the 0 axis, pixel points with pixel values of 1 are searched globally in the ROI, namely all white pixel values are searched in a binary image of a Sobel edge, the position of the white pixel value set is judged, and four highest vertical coordinate values in the set are obtained. The resulting four ordinates are the relative position of the 0 axis within the ROI.
S152, determining the position of the axis 0 in the region of interest according to the four highest longitudinal coordinate values in the position set;
s153, setting K to be 1, and extracting an upper envelope from top to bottom until a pixel point in the 0-axis range is identified;
s154, setting K to be-1, and extracting a lower envelope from bottom to top until a pixel point in a 0-axis range is identified;
and S155, returning to the step S153 until all pixel points in the region of interest are identified, and obtaining a brachial artery blood flow frequency spectrum change curve.
Specifically, after the 0-axis position is determined, the upper envelope and the lower envelope need to be respectively extracted in the Sobel gradient graph and the stitching needs to be completed. In order to accurately extract the upper and lower outer envelopes, the specific implementation of the envelope extraction algorithm is as follows. First, locking sets K to 1, extracts an upper envelope from top to bottom when K is 1, unlocks if a pixel point of 0-axis range is recognized when the envelope is extracted, and sets K to-1. When K is-1, the lower envelope is extracted from bottom to top, and when a pixel point of 0-axis range is identified, K is set to-K again. The identification, inversion, extraction and seamless splicing of the upper envelope and the lower envelope are finished. The method realizes the self-adaptive identification and extraction algorithm of different blood flow directions and obtains an accurate brachial artery blood flow maximum velocity curve.
As a preferred embodiment of the method, the ARX transfer function of the blood velocity spectrum is specifically a 12 th order ARX model, and the step of performing fitting analysis on the ARX transfer function of the blood velocity spectrum to obtain a transfer function characteristic includes:
and fitting and analyzing the ARX transfer function of the blood flow velocity spectrum based on a least square method, and extracting coefficient characteristics of the transfer function, horizontal and vertical coordinate characteristics of a zero point and pole horizontal and vertical coordinate characteristics.
According to data quality screening, 50 groups of coronary heart diseases (CAD) are established based on the blood flow spectrum curve, and 50 groups of system transfer function data are compared, namely blood flow velocity transfer functions from common carotid artery to brachial artery before, after EECP intervention; and finally, according to FPE precision judgment, the order is selected to be 12. Features such as coefficients of transfer functions and horizontal and vertical coordinates of poles-zero are extracted based on a 12 th-order ARX model, and the like, and the total number of the features is 73, and 219 features exist in three excitation states before, during and after EECP.
The ARX model can be expressed in a form of a linear regression equation, so that the parameters of the ARX model can be calculated by a least square method, and compared with other models, the ARX model is simple and convenient to calculate and wide in application; ARX is expressed as a difference equation:
where u is the system input, y is the system output, and e is white noise, which is the residual of the system. This system can also be expressed as:
A(q)y(t)＝B(q)u(t)+e(t)
wherein q is a transfer operator,
the output of the system, y (t), can be expressed as:
the ARX model parameters are calculated according to the formula, and the estimated value of the output quantity
Can be expressed as:
data vector of the model now introduced
And a parameter vector
The output quantity estimate may be expressed as a linear regression form as follows
Solving by adopting a least square method to minimize the error of the system; here, a loss function is introduced, i.e. the 2-norm of the system residual e is taken as an evaluation criterion of the system error:
wherein e can be represented as
Thus, the parameters of the system can be calculated using the following formula
The above equation is solved by using QR decomposition to obtain:
the above equation is called least squares estimation.
Further, as a preferred embodiment of the method, the step of processing the transfer function features by using an SVM classifier to complete classification of the blood flow spectrum signal specifically includes:
selecting the transfer function characteristics based on an LASSO regression algorithm to obtain selected pole horizontal and vertical coordinate characteristics;
classifying according to the selected zero-pole horizontal and vertical coordinate characteristics based on an SVM classifier to obtain a classification result;
and reflecting the corresponding blood flow spectrum signal classification according to the classification result.
According to the obtained blood flow velocity spectrum curve of the carotid artery and the brachial artery, a transfer function from the carotid artery to the peripheral brachial artery is established. Fitting analysis is carried out by adopting an ARX model, and the fitting degree of blood flow frequency spectrum curves before, in the middle and after EECP is more than 80% at 12 orders. Transfer function features computed in the model are used for classification. The difference in response of the transfer function characteristics to EECP excitation distinguishes CAD from control groups and is used for the aided diagnosis of disease.
As shown in fig. 2, a blood flow spectrum signal classification system includes:
the spectrum curve extraction module is used for acquiring an artery blood vessel ultrasonic image and extracting an artery blood flow spectrum change curve;
the transfer function building module is used for building an ARX transfer function of a blood velocity spectrum according to the arterial blood flow spectrum change curve;
the characteristic calculation module is used for performing fitting analysis on an ARX transfer function based on the blood flow velocity spectrum to obtain transfer function characteristics;
and the classification module is used for processing the transfer function characteristics by adopting an SVM classifier to finish the classification of the blood flow spectrum signals.
The contents in the above method embodiments are all applicable to the present system embodiment, the functions specifically implemented by the present system embodiment are the same as those in the above method embodiment, and the beneficial effects achieved by the present system embodiment are also the same as those achieved by the above method embodiment.
While the preferred embodiments of the present invention have been illustrated and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (7)
1. A blood flow spectrum signal classification method is characterized by comprising the following steps:
s1, acquiring an arterial blood vessel ultrasonic image and extracting an arterial blood flow frequency spectrum change curve;
s2, establishing an ARX transfer function of a blood velocity spectrum according to the arterial blood flow spectrum change curve;
s3, performing fitting analysis on the ARX transfer function of the blood flow velocity spectrum to obtain the transfer function characteristics;
and S4, processing the transfer function characteristics by adopting an SVM classifier to finish the classification of the blood flow spectrum signals.
2. The method for classifying a blood flow spectrum signal according to claim 1, wherein the step of acquiring an arterial blood vessel ultrasound image and extracting an arterial blood flow spectrum change curve specifically comprises:
s11, acquiring a carotid artery blood vessel ultrasonic image and a brachial artery blood vessel ultrasonic image;
s12, carrying out interference factor elimination processing, binarization processing, void filling processing and Sobel operator edge detection processing on the carotid artery blood vessel ultrasonic image to obtain a first Sobel gradient image;
s13, extracting the first Sobel gradient image to obtain a carotid blood flow frequency spectrum change curve;
s14, performing region-of-interest selection, binarization processing, cavity filling processing and Sobel operator edge detection processing on the brachial artery blood vessel ultrasonic image to obtain a second Sobel gradient image;
and S15, identifying the 0-axis position in the second Sobel gradient image, and locking the K value to search the upper envelope and the lower envelope to obtain the brachial artery blood flow frequency spectrum change curve.
3. The method for classifying blood flow spectrum signals according to claim 2, wherein the step of performing interference factor elimination processing, binarization processing, void filling processing and Sobel operator edge detection processing on the carotid artery blood vessel ultrasonic image to obtain a first Sobel gradient image specifically comprises:
s121, setting the pixel value of the interference factor in the carotid artery blood vessel ultrasonic image as background black to obtain a carotid artery blood flow frequency spectrum image with the interference factor removed;
s122, converting the carotid artery blood flow spectrum image after the interference factors are removed into a binary image;
s123, carrying out hole filling processing on the binary image by adopting closed operation to obtain a filled image;
and S124, respectively calculating gradients of the filled images in the horizontal direction and the vertical direction by adopting a Sobel operator and synthesizing to obtain a first Sobel gradient image.
4. The method for classifying blood flow spectrum signals according to claim 3, wherein the step of identifying the 0-axis position in the second Sobel gradient image and locking the K value to search the upper envelope and the lower envelope to obtain the brachial artery blood flow spectrum variation curve specifically comprises:
s151, searching pixel points with pixel values of 1 in the region of interest in the second Sobel gradient image in a global mode to obtain a position set of white pixel values;
s152, determining the position of the axis 0 in the region of interest according to the four highest longitudinal coordinate values in the position set;
s153, setting K to be 1, and extracting an upper envelope from top to bottom until a pixel point in the 0-axis range is identified;
s154, setting K to be-1, and extracting a lower envelope from bottom to top until a pixel point in a 0-axis range is identified;
and S155, returning to the step S153 until all pixel points in the region of interest are identified, and obtaining a brachial artery blood flow frequency spectrum change curve.
5. The method for classifying a blood flow spectrum signal according to claim 4, wherein the ARX transfer function of the blood flow velocity spectrum is a 12 th order ARX model, and the step of performing fitting analysis on the ARX transfer function of the blood flow velocity spectrum to obtain the transfer function characteristics comprises:
and fitting and analyzing the ARX transfer function of the blood flow velocity spectrum based on a least square method, and extracting coefficient characteristics of the transfer function, horizontal and vertical coordinate characteristics of a zero point and pole horizontal and vertical coordinate characteristics.
6. The method for classifying blood flow spectrum signals according to claim 5, wherein the step of processing the transfer function features by using the SVM classifier to complete the classification of the blood flow spectrum signals comprises:
selecting the transfer function characteristics based on an LASSO regression algorithm to obtain selected pole horizontal and vertical coordinate characteristics;
classifying according to the selected zero-pole horizontal and vertical coordinate characteristics based on an SVM classifier to obtain a classification result;
and reflecting the corresponding blood flow spectrum signal classification according to the classification result.
7. A blood flow spectral signal classification system, comprising:
the spectrum curve extraction module is used for acquiring an artery blood vessel ultrasonic image and extracting an artery blood flow spectrum change curve;
the transfer function building module is used for building an ARX transfer function of a blood velocity spectrum according to the arterial blood flow spectrum change curve;
the characteristic calculation module is used for performing fitting analysis on an ARX transfer function based on the blood flow velocity spectrum to obtain transfer function characteristics;
and the classification module is used for processing the transfer function characteristics by adopting an SVM classifier to finish the classification of the blood flow spectrum signals.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110928460.0A CN113724208B (en) | 2021-08-13 | 2021-08-13 | Blood flow spectrum signal classification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110928460.0A CN113724208B (en) | 2021-08-13 | 2021-08-13 | Blood flow spectrum signal classification method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113724208A true CN113724208A (en) | 2021-11-30 |
| CN113724208B CN113724208B (en) | 2023-06-06 |
Family
ID=78675731
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110928460.0A Active CN113724208B (en) | 2021-08-13 | 2021-08-13 | Blood flow spectrum signal classification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113724208B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114664414A (en) * | 2022-03-28 | 2022-06-24 | 中国人民解放军总医院第三医学中心 | Method and device for generating blood flow spectrum envelope curve and readable storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103049766A (en) * | 2012-12-27 | 2013-04-17 | 中国科学院遥感应用研究所 | Ultrasonic image renal artery blood flow spectrum signal curve classification method |
| CN104095656A (en) * | 2014-07-25 | 2014-10-15 | 声泰特（成都）科技有限公司 | Colorful blood flow imaging based on ultrasonic Doppler frequency spectrum and display method thereof |
| CN104182984A (en) * | 2014-09-01 | 2014-12-03 | 云南大学 | Method and system for rapidly and automatically collecting blood vessel edge forms in dynamic ultrasonic image |
| CN105105741A (en) * | 2015-07-15 | 2015-12-02 | 无锡海鹰电子医疗系统有限公司 | Envelope line extracting and feature point tracking method of pulse wave image |
| WO2018058606A1 (en) * | 2016-09-30 | 2018-04-05 | 深圳迈瑞生物医疗电子股份有限公司 | Method for displaying ultrasonic blood flow motion spectrum and ultrasonic imaging system thereof |
| CN107992452A (en) * | 2017-12-12 | 2018-05-04 | 北京动亮健康科技有限公司 | Calculate method, apparatus, storage medium and the equipment of central hemodynamics index |
| WO2019195156A1 (en) * | 2018-04-03 | 2019-10-10 | The Children' S Mercy Hospital | Systems and methods for detecting flow of biological fluids |
| CN111325202A (en) * | 2020-01-19 | 2020-06-23 | 华中科技大学同济医学院附属协和医院 | Ultrasonic Doppler-based color spectrogram frequency spectrum blood flow detection method |
-
2021
- 2021-08-13 CN CN202110928460.0A patent/CN113724208B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103049766A (en) * | 2012-12-27 | 2013-04-17 | 中国科学院遥感应用研究所 | Ultrasonic image renal artery blood flow spectrum signal curve classification method |
| CN104095656A (en) * | 2014-07-25 | 2014-10-15 | 声泰特（成都）科技有限公司 | Colorful blood flow imaging based on ultrasonic Doppler frequency spectrum and display method thereof |
| CN104182984A (en) * | 2014-09-01 | 2014-12-03 | 云南大学 | Method and system for rapidly and automatically collecting blood vessel edge forms in dynamic ultrasonic image |
| CN105105741A (en) * | 2015-07-15 | 2015-12-02 | 无锡海鹰电子医疗系统有限公司 | Envelope line extracting and feature point tracking method of pulse wave image |
| WO2018058606A1 (en) * | 2016-09-30 | 2018-04-05 | 深圳迈瑞生物医疗电子股份有限公司 | Method for displaying ultrasonic blood flow motion spectrum and ultrasonic imaging system thereof |
| CN107992452A (en) * | 2017-12-12 | 2018-05-04 | 北京动亮健康科技有限公司 | Calculate method, apparatus, storage medium and the equipment of central hemodynamics index |
| WO2019195156A1 (en) * | 2018-04-03 | 2019-10-10 | The Children' S Mercy Hospital | Systems and methods for detecting flow of biological fluids |
| CN111325202A (en) * | 2020-01-19 | 2020-06-23 | 华中科技大学同济医学院附属协和医院 | Ultrasonic Doppler-based color spectrogram frequency spectrum blood flow detection method |
Non-Patent Citations (3)
| Title |
|---|
| HAYIT GREENSPAN 等: "DOPPLER ECHOCARDIOGRAPHY FLOW-VELOCITY IMAGE ANALYSIS FOR PATIENTS WITH ATRIAL FIBRILLATION" * |
| JUERG TSCHIRREN 等: "Determination of the envelope function (maximum velocity curve) in Doppler ultrasound flow velocity diagrams" * |
| 税雪 等: "基于区域生长超声图像分割血管边界反映血管病变" * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114664414A (en) * | 2022-03-28 | 2022-06-24 | 中国人民解放军总医院第三医学中心 | Method and device for generating blood flow spectrum envelope curve and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113724208B (en) | 2023-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bai et al. | A bi-ventricular cardiac atlas built from 1000+ high resolution MR images of healthy subjects and an analysis of shape and motion | |
| CN110021037B (en) | Image non-rigid registration method and system based on generation countermeasure network | |
| US9968257B1 (en) | Volumetric quantification of cardiovascular structures from medical imaging | |
| WO2018120942A1 (en) | System and method for automatically detecting lesions in medical image by means of multi-model fusion | |
| Nambakhsh et al. | Left ventricle segmentation in MRI via convex relaxed distribution matching | |
| KR101599219B1 (en) | system and method for automatic registration of anatomic points in 3d medical images | |
| CN111340682A (en) | Method and system for converting medical image into different-style image by using deep neural network | |
| US9367924B2 (en) | Method and system for segmentation of the liver in magnetic resonance images using multi-channel features | |
| Zhang et al. | LU-NET: An improved U-Net for ventricular segmentation | |
| US8948484B2 (en) | Method and system for automatic view planning for cardiac magnetic resonance imaging acquisition | |
| CN111612756B (en) | Coronary artery specificity calcification detection method and device | |
| CN104732520A (en) | Cardio-thoracic ratio measuring algorithm and system for chest digital image | |
| WO2022086910A1 (en) | Anatomically-informed deep learning on contrast-enhanced cardiac mri | |
| CN112329844A (en) | Image object classification method and related device, equipment and storage medium | |
| Sani et al. | Automatic computation of left ventricular volume changes over a cardiac cycle from echocardiography images by nonlinear dimensionality reduction | |
| Almeida et al. | Left-atrial segmentation from 3-D ultrasound using B-spline explicit active surfaces with scale uncoupling | |
| US8064670B2 (en) | Analysis of multidimensional data | |
| CN113724208A (en) | Blood flow frequency spectrum signal classification method and system | |
| JP5532730B2 (en) | CNR measuring apparatus, method and computer program | |
| CN108597589B (en) | Model generation method, target detection method and medical imaging system | |
| Schwarz et al. | A deformable registration method for automated morphometry of MRI brain images in neuropsychiatric research | |
| CN109102509B (en) | Segmentation model training method and device and computer readable storage medium | |
| CN114049344A (en) | Image segmentation method, training method of model thereof, related device and electronic equipment | |
| CN111369598B (en) | Deep learning model training method and device, and application method and device | |
| Bernabé et al. | A self-optimized software tool for quantifying the degree of left ventricle hyper-trabeculation |
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 |