CN106570856A - Common carotid artery intima-media thickness measuring device and method combining level set segmentation and dynamic programming - Google Patents
Common carotid artery intima-media thickness measuring device and method combining level set segmentation and dynamic programming Download PDFInfo
- Publication number
- CN106570856A CN106570856A CN201610782555.5A CN201610782555A CN106570856A CN 106570856 A CN106570856 A CN 106570856A CN 201610782555 A CN201610782555 A CN 201610782555A CN 106570856 A CN106570856 A CN 106570856A
- Authority
- CN
- China
- Prior art keywords
- image
- dynamic programming
- roi
- lii
- estimated
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- 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
- A61B8/08—Detecting organic movements or changes, e.g. tumours, cysts, swellings
- A61B8/0858—Detecting organic movements or changes, e.g. tumours, cysts, swellings involving measuring tissue layers, e.g. skin, interfaces
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20028—Bilateral filtering
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20132—Image cropping
-
- 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
Abstract
The invention relates to medical instruments and image processing technologies, and aims to apply computer image processing technologies to common carotid artery ultrasonic image intima-media thickness measurement, avoid defects of low efficiency and instability existing in a manual measurement mode, provide quick and reliable IMT (Intima-Media Thickness) parameters by using a computer algorithm and provide an accurate basis for early diagnosis for cardiovascular diseases. According to the common carotid artery intima-media thickness measuring device and method combining level set segmentation and dynamic programming, processing is performed on a received ultrasonic image by a computer, wherein the computer is provided with the following modules: 1) an image cutting module, 2) an ROI (Region of Interest) extraction module, 3) an image dynamic stretching module which stretches the image grayscale range to 0-255 in a unified manner, 4) an image filtering module, 5) an LII (Lumen-Intima Interface) estimation module, 6) an MAI (Media-Advantia Interface) estimation module, 7) a dynamic programming adjustment and interface estimation module, and 8) a post-processing module. The common carotid artery intima-media thickness measuring device and method are mainly applied to designing and manufacturing of medical instruments.
Description
Technical field
The present invention relates to medical apparatus and instruments, image processing techniquess, specifically, are related to based on level set and dynamic programming intra-arterial
Media thickness measurement apparatus and algorithm.
Background technology
Cardiovascular and cerebrovascular disease (Cardiovascular diseases, CVDs) is the primary disease of human body health,
Jing becomes the killer of human life first.And the modal diseased region of arteriosclerosis is exactly common carotid artery (Common
Carotid Artery, CCA), the lipid in blood with cardiovascular and cerebrovascular disease development in intravascular wall deposition so that interior
The thickening hypertrophy of middle membrane structure (Intima-Media Complex, IMC), tremulous pulse narrows, Internal-media thickness (Intima-Media
Thickness, IMT) if increase is to a certain extent, blood flow rate will slow down, and make blood supply in brain not smooth, and brain function is also received
To impact.Therefore, the early lesion of cardiovascular and cerebrovascular disease shows on interior medial thickening that Internal-media thickness is prediction cardiovascular and cerebrovascular vessel
The important indicator of disease severity.Clinically, Internal-media thickness is typically doctor is carried out to common carotid artery ultrasonoscopy
Manual fixed point or border delineation show that subjectivity is strong, becomes more dependent on personal experience.Therefore, the present invention proposes one kind and combines water
Flat collection segmentation and the carotid ultrasound image Internal-media thickness Measurement Algorithm of dynamic programming, being replaced using computer generation manually carries out side
IMT relevant parameters are split and calculated to limitans automatically, to reduce artificial time-consuming and workload, and reduces measuring difference, is heart and brain blood
Pipe disease early diagnosiss provide accurate foundation.
Level Set Method based on geometry deformation model is proposed by Osher and Sethian in nineteen eighty-two.Using level set point
Geometric curve evolution model theory represented by segmentation method is applied to the ultimate principle of image segmentation field:Using velocity function
F (include curve deformation power, such as Curvature-driven power or normal direction constant power or), image spatial dimension to level set function
Develop, the position that causing the closed curve described by the zero level collection of level set function can be located in target profile curve is stopped
Only.Most important in the thought of Level Set Method is exactly Hamilton-Jacobi equations, and it carries out base for the implicit surface of motion
It is to be embedded into curve or curved surface as zero level collection high one-dimensional in the ultimate principle of numerical solution the method for the equation of time
Level set function in, the curve of low-dimensional or the evolutionary process of curved surface are expressed by the function of a more higher-dimension.Use level set
The zero level set function of function represents that the core concept of the description method of closed curve is:Closed curve the C={ (x for developing
(p), y (p)), p ∈ [0,1] } it is embedded into 2D level set functions φ (x, y), with the zero level collection of level set function implicitly
Represent closed curve C:C={ (x, y), φ (x, y)=0 }.The now normal vector of curveTangent vectorAnd have in t
φ (C (p, t), t)=0 (1)
Both sides derivation is obtained
The curve evolvement equation and normal vector of substitution formula (1)Level set movements equation is obtained
Wherein F is the velocity function that level set function develops, and similarly there is also different drilling according to the difference of its value
Change mode.Constantly developed level set function by the effect of velocity function F, the profile that last zero level set function is formed is just
It is the final form of curve evolvement.
Dynamic programming (Dynamic Programming) is a kind of efficient optimization method, can solve function argument
Optimization problem in the case of non-three phase pass.It is critical only that for dynamic programming techniques defines a cost function, the cost function bag
Include the constraints such as edge strength, border be smooth, make cost function reach the minimum border that can be extracted from left to right it by challenge
A series of simple subproblems are decomposed into, and the solution of problem is obtained by the method for iteration.For following this problem:
h(x1,x2,x3,x4)=h1(x1,x2)+h2(x2,x3)+h3(x3,x4) (4)
Solve (x1,x2,x3,x4) so that h (x1,x2,x3,x4) obtain maximum.
Following four step can be divided into when solving this problem using dynamic programming algorithm:
The first step, for each x2Calculate f1(x2) maximum and preserve corresponding x1With f1(x2) value, i.e.,
Second step, for each x3Calculate f2(x3) maximum and preserve corresponding x2With f2(x3) value, i.e.,
3rd step, for each x4Calculate f3(x4) maximum and preserve corresponding x3With f3(x4) value, i.e.,
4th step, selects x4So that f3(x4) value reaches maximum, and selected and the f by " backstepping method "3(x4) corresponding x3With
f2(x3) corresponding x2And with f1(x2) corresponding x1。
Finally solve the (x for meeting condition1,x2,x3,x4)。
The advantage that dynamic programming is solved is that calculating speed is fast.For containing N number of independent variable (x1,x2,...,xN), each from
Variable has the h (x of K kind values1,x2,...,xN) optimization problem, the amount of calculation of dynamic programming algorithm is (N-1) K2+N.For same
The problem of control gauge mould, if the amount of calculation solved with greedy algorithm is KN。
The content of the invention
To overcome the deficiencies in the prior art, it is contemplated that computer image processing technology is applied in carotid ultrasound ripple
In the measurement of image Internal-media thickness, it is to avoid the inefficient and unstable defect that manual measurement mode is present, calculated using computer
Method provides quick, reliable IMT parameters, and for cardiovascular and cerebrovascular disease early diagnosiss accurate foundation is provided.The skill that the present invention is adopted
Art scheme is that level set dynamic programming Carotid intima-media thickness measuring method comprises the steps:
1) image cropping:Cutting removes non-test message part in ultrasonoscopy, an artery-sparing part;
2) region of interest ROI (Region of Interest) is extracted:Automatically extract ROI;
3) image dynamic tensile:The unification of gradation of image scope is stretched to into 0-255;
4) image filtering:Noise is removed using bilateral filtering algorithm, while retaining image edge information;
5) tube chamber-intima boundary LII (Lumen-Intima Interface) is estimated:Inner membrance is avoided using morphological operation
The interference of defect so that the LII of estimation is closer to being really border;
6) film-epicardial border MAI (Media-Advantia Interface) in estimating:Estimated according to gradient map information
MAI;
7) border is estimated in dynamic programming adjustment:Using the weighted sum of Grad and curvature value as the energy letter of dynamic programming
Number, is iterated calculating and obtains membrane boundary characteristic;
8) post processing:IMT is calculated and assessed, the IMT values of automatic measurement are differentiated, it is to avoid measurement result is without actual
Meaning.
Extract ROI to comprise the concrete steps that, directly image border described using full curve first, at the same time using image
Information is defined to energy functional, and the independent variable of energy functional contains boundary profile curve;Then dynamic Euler- is adopted
The form of Lagrange equations obtains the equation of the corresponding curve evolvement of this energy functional;Finally adopt Level Set Method pair
The evolutionary process for declining most fast direction along Energy Simulation initial curve is simulated, bent to obtain optimal boundary profile
Line, extracts ROI section.
Image dynamic tensile comprises the concrete steps that the gray level that will be above m is compressed, until it is compressed to high grade grey level
In narrower range, and the gray level that will be less than m is compressed so as to be present in the narrower range of low gray level, in the middle of stretching
Gray level, strengthen each several part image contrast so that image possesses the high contrast of comparison;Stretching formula is as follows:
Wherein, s represents the gray value of output image, and r represents input picture gray value, E be one can be oblique with control function
The variable of rate, m as a threshold value can by user according to image procossing in terms of be actually needed and carry out parameter setting.
Bilateral filtering formula is as follows:
Wherein,For the normaliztion constant of bilateral filtering, x, ξ represents pixel, f (x)
The gray value of x pixels is represented, wave filter is the matrix of size 10 × 10,Represent picture
The gray scale similarity of plain ξ and x,Represent the Distance conformability degree of pixel ξ and x, σ1It is based on gray scale phase
The Gauss standard calculated like property is poor, and the Gaussian function standard deviation just calculated based on distance similarity is σ2。
Estimate LII steps specifically, the gray value of ROI is normalized in the range of interval [0,1], the number of normalized function
Learn expression formula as follows:
Wherein, g (x) be gray scale actual value, fminAnd fmaxIt is respectively the minima and maximum of gray value in ROI;
The rectangular histogram obtained after normalization is made up of 3 peaks and 2 paddy.The gray value for assuming 2 paddy minimum points is T1With
T2, vertically to search for from top to bottom in the roi, concrete search step is as follows:
(1) j=1 is made;By i1Increase to M from 1 to scan for until f (i1,j)≥T1;(i1, j) i.e. estimated LII
Point;
(2) it is j+1 to increase j, i.e., the next column in ROI is searched for until f (i2,j)≥T1;
(3) if i2>i1, by (i1+ 1, j) as the new point estimated on LII, if i2<i1, by (i1- 1, j) conduct is estimated
A new point on meter LII, otherwise, by (i1, j) as new estimation point;The vertical coordinate bit of latest estimated point is installed
For i1;
(4) if j=N, the LII of output estimation;Otherwise, return to step (2).
Membrane boundary is comprised the concrete steps that in dynamic programming adjustment, and the factor according to of both gradient with curvature is following to set up
Equation:
Wherein, g (xk) represent pixel xkThe Grad of position, c (xk-1,xk,xk+1) represent by (xk-1,xk,xk+1) three points
It is linked to be the amount of curvature of broken line, λ is a negative constant.Preferably border has Grad big and the characteristics of little curvature, accordingly, it would be desirable to
Solve (x1,x2,...,xN) cause L (x1,x2,...,xN) functional value reaches maximum, discounted method using eight neighborhood, by broken line
Each point is up or certain pixel distance that moves down so that L (x1,x2,...,xN) reach maximum, that is, obtain most reasonable
Border.
Level set dynamic programming Carotid intima-media thickness measurement apparatus, by ultrasound image acquisition device and calculating mechanism
Into, computer to processing after the ultrasonoscopy that receives, it is provided with such as lower module on computer:
1) image cropping module:Cutting removes non-test message part in ultrasonoscopy, an artery-sparing part;
2) ROI modules are extracted:Automatically extract ROI;
3) image dynamic tensile module:The unification of gradation of image scope is stretched to into 0-255;
4) image filtering module:Noise is removed using bilateral filtering algorithm, while retaining image edge information;
5) LII modules are estimated:The interference of inner membrance defect is avoided using morphological operation so that the LII of estimation closer to
Really it is border;
6) MAI modules are estimated:MAI is estimated according to gradient map information;
7) boundary module is estimated in dynamic programming adjustment:Using the weighted sum of Grad and curvature value as the energy of dynamic programming
Function, is iterated calculating and obtains membrane boundary characteristic;
8) post-processing module:IMT is calculated and assessed, the IMT values of automatic measurement is differentiated, it is to avoid measurement result does not have
Practical significance.
Of the invention the characteristics of and beneficial effect are:
Level set algorithm utilizes velocity function, and level set function is developed in the spatial dimension of image, causes level set letter
The position that closed curve described by several zero level collection can be located in target profile curve stops.Level set algorithm can very certainly
Right process interface change in topology, it is easy to solve higher-dimension problem.Present invention introduces Level Set Method extracts ROI, it is subsequent step
Initial pictures are provided, by morphological dilations etching operation, make image result relatively reliable, with reference to dynamic programming algorithm, with ladder
The weighted sum of angle value and curvature value accurately have adjusted two borders of LII and MAI as its energy function, make Internal-media thickness
Measurement result is more accurately and reliably.The measurement of carotid ultrasound image Internal-media thickness in strong support of the present invention clinic, be
The further optimized development of IMT computer aided measurement technologies provides reference, is good to the mode of expert's manual measurement
Supplement.
Area-of-interest is partitioned into using Level Set Method in the present invention, final border is entered with reference to dynamic programming algorithm
The accurate adjustment of row.In the image library test given, the present invention can efficiently accomplish in common carotid artery the segmentation of middle membrane boundary and have
Effect calculates IMT values, with preferable theoretical and use value.
Description of the drawings:
Fig. 1 algorithm flow charts;
The initial carotid ultrasound images of Fig. 2;
Fig. 3 cuttings simultaneously extract the image of ROI;
The image result of Fig. 4 dynamic tensiles;
Fig. 5 filters wavefront image;
Fig. 6 filtered images;
The initial LII borders of Fig. 7;
Initial LII, MAI borders of Fig. 8;
Final LII, MAI border after Fig. 9 dynamic programmings adjustment.
Specific embodiment
The present invention is adopted the following technical scheme that:
1) image cropping.The relevant information of patient and ultrasonic instrument is distributed with around initial ultrasound image, these information
Presence can affect follow-up image processing step, it is therefore desirable to cutting removes these parts, an artery-sparing part.
2) ROI is extracted.Automatically extracting ROI can avoid that manual segmentation ROI's is loaded down with trivial details, realize the purpose of automatic measurement IMT.
3) image dynamic tensile.The unification of gradation of image scope is stretched to into 0-255, makes image outline apparent.
4) image filtering.Ultrasonoscopy is become neck and is moved with the advantage of real-time, repeatability, Noninvasive and low cost
The first-selected imaging mode that arteries and veins is checked.Bilateral filtering algorithm can preferably retain image edge information while noise is removed.
5) LII is estimated.Morphological operation avoids the interference of inner membrance defect so that the LII of estimation is closer to being really side
Boundary.
6) MAI is estimated.Estimate more accurate than simple translation according to the MAI that gradient map information is estimated.
7) border is estimated in dynamic programming adjustment.Using the weighted sum of Grad and curvature value as the energy letter of dynamic programming
Number, the result of iteration more conforms to the membrane boundary characteristic of common carotid artery.
8) post processing.Calculate and assess IMT.The IMT values of automatic measurement are differentiated, it is to avoid measurement result is without actual
Meaning.
Below in conjunction with the accompanying drawings the invention will be further described with example.
1) image cropping.Under conditions of ensureing that distal vessels wall is present, 320 × 300 pictures of lower section in ultrasonoscopy are intercepted
The region of plain size.
2) ROI is extracted.
First have to that directly image border is described using full curve, it is at the same time general to a certain energy using image information
Letter is defined, and the independent variable of energy functional contains boundary profile curve.Then using dynamic Euler-Lagrange equations
Form obtains the equation of the corresponding curve evolvement of this energy functional;Finally using Level Set Method to along at the beginning of Energy Simulation
Beginning curve declines the evolutionary process in most fast direction and is simulated, and to obtain optimal boundary profile curve, extracts ROI portions
Point.
3) picture contrast dynamic tensile.
Directly greyscale transformation is to make picture superposition using the tonal range for changing selected image, and contrast is relatively small
Image therefore level more enriches, hence improve the perception of vision, be finally reached the effect of image enhaucament.Will be above m's
Gray level is compressed, and until it is compressed in the narrower range of high grade grey level, and the gray level that will be less than m is compressed, and makes
It is present in the narrower range of low gray level, stretches middle gray level, can strengthen the contrast of each several part image, so that figure
As possessing the high contrast of comparison.
Wherein, s represents the gray value of output image, and r represents input picture gray value, E be one can be oblique with control function
The variable of rate.M as a threshold value can by user according to image procossing in terms of be actually needed and carry out parameter setting.
4) bilateral filtering.
For image f (x), g (x) is obtained based on the grey similarity gaussian filtering of pixel, mathematic(al) representation is:
Wherein,For the normaliztion constant of grey similarity gaussian filtering, x, ξ represents pixel,
F (x) represents the gray value of x pixels,Represent the gray scale similarity of pixel ξ and x.ξ with
The gray value difference of x is less, and gray scale similarity s is bigger.
Image f (x) obtains g (x) through the gaussian filtering based on pixel distance similarity, and mathematic(al) representation is:
Wherein,For the normaliztion constant of the gaussian filtering of distance similarity,Represent the Distance conformability degree of pixel ξ and x.The distance of ξ and x is nearer, and c values are bigger.
Gray similarity and distance similarity these two aspects, obtain bilateral filtering formula as follows:
Wherein,For the normaliztion constant of bilateral filtering, x represents pixel, wave filter
For the matrix of size 10 × 10, wherein, the Gauss standard difference calculated based on grey similarity is σ1=0.1, based on apart from phase
It is σ like the Gaussian function standard deviation that property is just calculated2=3.
5) LII is estimated.
The gray value of ROI is normalized in the range of interval [0,1], the mathematic(al) representation of normalized function is as follows:
Wherein, g (x) be gray scale actual value, fminAnd fmaxIt is respectively the minima and maximum of gray value in ROI.
The rectangular histogram obtained after normalization is made up of 3 peaks and 2 paddy.The gray value for assuming 2 paddy minimum points is T1With
T2, vertically to search for from top to bottom in the roi, concrete search step is as follows:
(1) j=1 is made;By i1Increase to M from 1 to scan for until f (i1,j)≥T1;(i1, j) i.e. estimated LII
Point;
(2) it is j+1 to increase j, i.e., the next column in ROI is searched for until f (i2,j)≥T1。
(3) if i2>i1, by (i1+ 1, j) as the new point estimated on LII, if i2<i1, by (i1- 1, j) conduct is estimated
A new point on meter LII, otherwise, by (i1, j) as new estimation point;The vertical coordinate bit of latest estimated point is installed
For i1。
(4) if j=N, the LII of output estimation;Otherwise, return to step (2).
6) MAI is estimated.
LII is carried out after morphological dilations corrosion, if the LII of estimation 17 pixel distances that move down are obtained into rough
MAI estimates.
7) middle membrane boundary in dynamic programming adjustment.
Below equation is set up according to gradient and curvature both sides factor:
Wherein, g (xk) represent pixel xkThe Grad of position, c (xk-1,xk,xk+1) represent by (xk-1,xk,xk+1) three points
It is linked to be the amount of curvature of broken line, λ is a negative constant.Preferably border has Grad big and the characteristics of little curvature, accordingly, it would be desirable to
Solve (x1,x2,...,xN) cause L (x1,x2,...,xN) functional value reaches maximum, discounted method using eight neighborhood, by broken line
Each point is up or certain pixel distance that moves down so that L (x1,x2,...,xN) reach maximum, that is, obtain most reasonable
Border.
Claims (6)
1. a kind of level set dynamic programming Carotid intima-media thickness measuring method, is characterized in that, comprise the steps:
1) image cropping:Cutting removes non-test message part in ultrasonoscopy, an artery-sparing part;
2) region of interest ROI (Region of Interest) is extracted:Automatically extract ROI;
3) image dynamic tensile:The unification of gradation of image scope is stretched to into 0-255;
4) image filtering:Noise is removed using bilateral filtering algorithm, while retaining image edge information;
5) tube chamber-intima boundary LII (Lumen-Intima Interface) is estimated:Inner membrance defect is avoided using morphological operation
Interference so that the LII of estimation is closer to being really border;
6) film-epicardial border MAI (Media-Advantia Interface) in estimating:MAI is estimated according to gradient map information;
7) border is estimated in dynamic programming adjustment:Weighted sum using Grad and curvature value is entered as the energy function of dynamic programming
Row iteration is calculated and obtains membrane boundary characteristic;
8) post processing:IMT is calculated and assessed, the IMT values of automatic measurement are differentiated, it is to avoid measurement result is without actual meaning
Justice.
2. level set dynamic programming Carotid intima-media thickness measuring method as claimed in claim 1, is characterized in that, extract
ROI comprises the concrete steps that, directly image border described using full curve first, at the same time general to energy using image information
Letter is defined, and the independent variable of energy functional contains boundary profile curve;Then using dynamic Euler-Lagrange equations
Form obtains the equation of the corresponding curve evolvement of this energy functional;Finally using Level Set Method to along at the beginning of Energy Simulation
Beginning curve declines the evolutionary process in most fast direction and is simulated, and to obtain optimal boundary profile curve, extracts ROI portions
Point.
3. level set dynamic programming Carotid intima-media thickness measuring method as claimed in claim 1, is characterized in that, image
Dynamic tensile comprises the concrete steps that the gray level that will be above m is compressed, until it is compressed in the narrower range of high grade grey level,
And the gray level that will be less than m is compressed so as to be present in the narrower range of low gray level, middle gray level is stretched, increased
The contrast of strong each several part image, so that image possesses the high contrast of comparison;Stretching formula is as follows:
Wherein, s represents the gray value of output image, and r represents input picture gray value, E be one can be with control function slope
Variable, m as a threshold value can by user according to image procossing in terms of be actually needed and carry out parameter setting.
4. level set dynamic programming Carotid intima-media thickness measuring method as claimed in claim 1, is characterized in that, bilateral
Filtering Formula is as follows:
Wherein,For the normaliztion constant of bilateral filtering, x, ξ represents pixel, and f (x) represents x
The gray value of pixel, wave filter is the matrix of size 10 × 10,Represent pixel ξ and x
Gray scale similarity,Represent the Distance conformability degree of pixel ξ and x, σ1It is based on grey similarity meter
The Gauss standard for calculating is poor, and the Gaussian function standard deviation just calculated based on distance similarity is σ2。
5. level set dynamic programming Carotid intima-media thickness measuring method as claimed in claim 1, is characterized in that, estimate
LII steps specifically, the gray value of ROI are normalized in the range of interval [0,1], and the mathematic(al) representation of normalized function is such as
Under:
Wherein, g (x) be gray scale actual value, fminAnd fmaxIt is respectively the minima and maximum of gray value in ROI;
The rectangular histogram obtained after normalization is made up of 3 peaks and 2 paddy.The gray value for assuming 2 paddy minimum points is T1And T2,
Vertically search for from top to bottom in ROI, concrete search step is as follows:
(1) j=1 is made;By i1Increase to M from 1 to scan for until f (i1,j)≥T1;(i1, j) i.e. estimated LII starting points;
(2) it is j+1 to increase j, i.e., the next column in ROI is searched for until f (i2,j)≥T1;
(3) if i2> i1, by (i1+ 1, j) as the new point estimated on LII, if i2< i1, by (i1- 1, j) as estimation
A new point on LII, otherwise, by (i1, j) as new estimation point;The vertical coordinate position of latest estimated point is set to
i1;
(4) if j=N, the LII of output estimation;Otherwise, return to step (2).
Membrane boundary is comprised the concrete steps that in dynamic programming adjustment, is set up with lower section according to gradient and curvature both sides factor
Journey:
Wherein, g (xk) represent pixel xkThe Grad of position, c (xk-1,xk,xk+1) represent by (xk-1,xk,xk+1) three points are linked to be
The amount of curvature of broken line, λ is a negative constant.Preferably border has Grad big and the characteristics of little curvature, accordingly, it would be desirable to solve
(x1,x2,...,xN) cause L (x1,x2,...,xN) functional value reaches maximum, discounted method using eight neighborhood, by each on broken line
Point is up or certain pixel distance that moves down so that L (x1,x2,...,xN) reach maximum, that is, obtain most rational side
Boundary.
6. a kind of level set dynamic programming Carotid intima-media thickness measurement apparatus, is characterized in that, be filled by ultrasound image acquisition
Put and computer is constituted, computer is provided with such as lower module to processing after the ultrasonoscopy that receives on computer:
1) image cropping module:Cutting removes non-test message part in ultrasonoscopy, an artery-sparing part;
2) ROI modules are extracted:Automatically extract ROI;
3) image dynamic tensile module:The unification of gradation of image scope is stretched to into 0-255;
4) image filtering module:Noise is removed using bilateral filtering algorithm, while retaining image edge information;
5) LII modules are estimated:The interference of inner membrance defect is avoided using morphological operation so that the LII of estimation is closer to being really
Border;
6) MAI modules are estimated:MAI is estimated according to gradient map information;
7) boundary module is estimated in dynamic programming adjustment:Using the weighted sum of Grad and curvature value as the energy letter of dynamic programming
Number, is iterated calculating and obtains membrane boundary characteristic;
8) post-processing module:IMT is calculated and assessed, the IMT values of automatic measurement are differentiated, it is to avoid measurement result is without actual
Meaning.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610782555.5A CN106570856A (en) | 2016-08-31 | 2016-08-31 | Common carotid artery intima-media thickness measuring device and method combining level set segmentation and dynamic programming |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610782555.5A CN106570856A (en) | 2016-08-31 | 2016-08-31 | Common carotid artery intima-media thickness measuring device and method combining level set segmentation and dynamic programming |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106570856A true CN106570856A (en) | 2017-04-19 |
Family
ID=58532411
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610782555.5A Pending CN106570856A (en) | 2016-08-31 | 2016-08-31 | Common carotid artery intima-media thickness measuring device and method combining level set segmentation and dynamic programming |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106570856A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108961278A (en) * | 2018-06-20 | 2018-12-07 | 深圳市旭东数字医学影像技术有限公司 | The method and its system of abdominal wall muscle segmentation based on image data |
CN109493383A (en) * | 2018-11-23 | 2019-03-19 | 深圳市威尔德医疗电子有限公司 | The measurement method of Internal-media thickness, server and storage medium in ultrasound image |
CN109615595A (en) * | 2018-12-03 | 2019-04-12 | 中国石油大学(华东) | A kind of level set SAR oil spilling extracting method based on bilateral filtering |
CN109919953A (en) * | 2019-01-21 | 2019-06-21 | 深圳蓝韵医学影像有限公司 | Method, system and the equipment of carotid intimal medial thickness measurement |
CN110517263A (en) * | 2019-09-02 | 2019-11-29 | 青岛海信医疗设备股份有限公司 | Determine the method, apparatus and storage medium of Internal-media thickness |
CN113974688A (en) * | 2021-09-18 | 2022-01-28 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging method and ultrasonic imaging system |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102446357A (en) * | 2011-11-23 | 2012-05-09 | 浙江工商大学 | Level set SAR (Synthetic Aperture Radar) image segmentation method based on self-adaptive finite element |
CN102800089A (en) * | 2012-06-28 | 2012-11-28 | 华中科技大学 | Main carotid artery blood vessel extraction and thickness measuring method based on neck ultrasound images |
CN103093473A (en) * | 2013-01-25 | 2013-05-08 | 北京理工大学 | Multi-target picture segmentation based on level set |
CN103593848A (en) * | 2013-11-25 | 2014-02-19 | 深圳市恩普电子技术有限公司 | Ultrasonic endocardium tracing method |
CN103679728A (en) * | 2013-12-16 | 2014-03-26 | 中国科学院电子学研究所 | Water area automatic segmentation method of SAR image of complicated terrain and device |
CN104665872A (en) * | 2014-12-29 | 2015-06-03 | 深圳开立生物医疗科技股份有限公司 | Ultrasonic image-based carotid intima-media thickness measuring method and device |
-
2016
- 2016-08-31 CN CN201610782555.5A patent/CN106570856A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102446357A (en) * | 2011-11-23 | 2012-05-09 | 浙江工商大学 | Level set SAR (Synthetic Aperture Radar) image segmentation method based on self-adaptive finite element |
CN102800089A (en) * | 2012-06-28 | 2012-11-28 | 华中科技大学 | Main carotid artery blood vessel extraction and thickness measuring method based on neck ultrasound images |
CN103093473A (en) * | 2013-01-25 | 2013-05-08 | 北京理工大学 | Multi-target picture segmentation based on level set |
CN103593848A (en) * | 2013-11-25 | 2014-02-19 | 深圳市恩普电子技术有限公司 | Ultrasonic endocardium tracing method |
CN103679728A (en) * | 2013-12-16 | 2014-03-26 | 中国科学院电子学研究所 | Water area automatic segmentation method of SAR image of complicated terrain and device |
CN104665872A (en) * | 2014-12-29 | 2015-06-03 | 深圳开立生物医疗科技股份有限公司 | Ultrasonic image-based carotid intima-media thickness measuring method and device |
Non-Patent Citations (1)
Title |
---|
QIANG LI,WEI ZHANG,ETC.: ""An Improved Approach for Accurate and Efficient Measurement of Common Carotid Artery Intima-Media Thickness in Ultrasound Images"", 《BIOMED RESEARCH INTERNATIONAL》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108961278A (en) * | 2018-06-20 | 2018-12-07 | 深圳市旭东数字医学影像技术有限公司 | The method and its system of abdominal wall muscle segmentation based on image data |
CN108961278B (en) * | 2018-06-20 | 2023-01-31 | 深圳市旭东数字医学影像技术有限公司 | Method and system for abdominal wall muscle segmentation based on image data |
CN109493383A (en) * | 2018-11-23 | 2019-03-19 | 深圳市威尔德医疗电子有限公司 | The measurement method of Internal-media thickness, server and storage medium in ultrasound image |
CN109493383B (en) * | 2018-11-23 | 2022-02-11 | 深圳市威尔德医疗电子有限公司 | Method for measuring intima-media thickness in ultrasonic image, server and storage medium |
CN109615595A (en) * | 2018-12-03 | 2019-04-12 | 中国石油大学(华东) | A kind of level set SAR oil spilling extracting method based on bilateral filtering |
CN109615595B (en) * | 2018-12-03 | 2019-07-23 | 中国石油大学(华东) | A kind of level set SAR oil spilling extracting method based on bilateral filtering |
CN109919953A (en) * | 2019-01-21 | 2019-06-21 | 深圳蓝韵医学影像有限公司 | Method, system and the equipment of carotid intimal medial thickness measurement |
CN110517263A (en) * | 2019-09-02 | 2019-11-29 | 青岛海信医疗设备股份有限公司 | Determine the method, apparatus and storage medium of Internal-media thickness |
CN113974688A (en) * | 2021-09-18 | 2022-01-28 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging method and ultrasonic imaging system |
CN113974688B (en) * | 2021-09-18 | 2024-04-16 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging method and ultrasonic imaging system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106570856A (en) | Common carotid artery intima-media thickness measuring device and method combining level set segmentation and dynamic programming | |
US10325369B2 (en) | Method and system for analyzing blood flow condition | |
CN102163326B (en) | Method for automatic computerized segmentation and analysis on thickness uniformity of intima media of carotid artery blood wall in sonographic image | |
CN102890823B (en) | Motion object outline is extracted and left ventricle image partition method and device | |
CN102243759B (en) | Three-dimensional lung vessel image segmentation method based on geometric deformation model | |
US8218845B2 (en) | Dynamic pulmonary trunk modeling in computed tomography and magnetic resonance imaging based on the detection of bounding boxes, anatomical landmarks, and ribs of a pulmonary artery | |
US10134143B2 (en) | Method for acquiring retina structure from optical coherence tomographic image and system thereof | |
CN111640120A (en) | Pancreas CT automatic segmentation method based on significance dense connection expansion convolution network | |
CN106485203A (en) | Carotid ultrasound image Internal-media thickness measuring method and system | |
CN110008992B (en) | Deep learning method for prostate cancer auxiliary diagnosis | |
CN108109151B (en) | Method and device for segmenting ventricle of echocardiogram based on deep learning and deformation model | |
CN104545999B (en) | Method and device for measuring bladder volume through ultrasound images | |
CN106570871A (en) | Fuzzy C mean value carotid ultrasonic image intima-media thickness measuring method and system | |
CN113781640A (en) | Three-dimensional face reconstruction model establishing method based on weak supervised learning and application thereof | |
CN115830016B (en) | Medical image registration model training method and equipment | |
CN112288794B (en) | Method and device for measuring blood vessel diameter of fundus image | |
CN105894498A (en) | Optical coherent image segmentation method for retina | |
Gharleghi et al. | Deep learning for time averaged wall shear stress prediction in left main coronary bifurcations | |
Almeida et al. | Left-atrial segmentation from 3-D ultrasound using B-spline explicit active surfaces with scale uncoupling | |
CN103700068B (en) | A kind of method that in CTA image, liver and blood vessel are split simultaneously | |
CN104751457A (en) | Novel variational energy based liver partition method | |
CN104546000A (en) | Shape feature-based ultrasonic image bladder volume measuring method and device | |
Qin et al. | Predicting tongue shapes from a few landmark locations | |
CN110689080B (en) | Planar atlas construction method of blood vessel structure image | |
Huang et al. | Adoption of snake variable model-based method in segmentation and quantitative calculation of cardiac ultrasound medical images |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170419 |