CN103606145A - Method for segmenting aortic-valve ultrasound image sequence based on interframe-shape-constraint GCV model - Google Patents

Abstract

The invention discloses a method for segmenting an aortic-valve ultrasound image sequence based on an interframe-shape-constraint GCV model. The method includes the following steps: acquiring a group of ultrasound images and carrying out Wiener filtering; calculating a gradient vector flow field of each image frame by frame and adding the gradient vector flow fields to a CV model as energy constraints and obtaining the GCV model; through defining an initial constraint shape and adding the initial constraint shape into the GCV model as an energy constraint item and then minimizing an energy functional so that a segmentation result of a first frame of image is obtained; and carrying out filtering of a rolling sphere method on the aortic-valve segmentation result of an adjacent previous-frame image and adding the obtained result into the GCV model as an energy constraint item so that the segmentation result of a current frame is obtained through calculation. The method for segmenting the aortic-valve ultrasound image sequence based on the interframe-shape-constraint GCV model is operated on short-axis images of ultrasonic cardiograms so that not only is work load of doctors reduced, a problem that serious overflow in segmentation of the aortic-valve ultrasound images in the prior art is also solved. The segmentation result is extremely close to the result of manual segmentation. The method enables the aortic valve to be segmented in a simple and highly-efficient manner.

Description

GCV model based on interframe shape constraining is cut apart aorta petal ultrasonic image sequence method

Technical field

The present invention relates to Medical Ultrasonic Image Segmentation method, a kind of GCV model based on interframe shape constraining is cut apart aorta petal ultrasonic image sequence method specifically.

Background technology

In China, aortic valve class disease be a kind of the most common be also complicated, the most dangerous angiocardiopathy, the healthy of the mankind in serious harm.Aorta petal is positioned at left ventricular outflow tract end and sustainer intersection, and it acts on just as one " one-way cock ", prevents blood backflow, guarantees that cardiac pumping function is normally good.Due to B ultrasonic safety without wound, simple cheap, can repeat, in clinical diagnosis, be widely used shape and the position of Echocardiographic Observation aorta petal at present.In its ultrasonoscopy, can assist doctor's clinical diagnosis accurately cutting apart not only of aorta petal, is also the basis that image guides non-intervention valve class operation simultaneously.But, because its ultrasonoscopy has low contrast, exists the ultrasonic and aorta petal of a large amount of spots constantly to open and close the feature of motion, at present in clinical diagnosis, mainly to rely on the doctor who has a large amount of clinical experiences manually to be cut apart one by one aorta petal ultrasonoscopy, this has not only increased doctor's workload greatly, and for the doctor who is relatively short of for clinical experience, manually cutting apart is also the work being relatively not easy.

For solving all inconvenience of manually cutting apart existence, there have been many scholars to propose the automatic or semi-automatic method of cutting apart of multiple ultrasonoscopy at home and abroad at present.As at home, 2005, the people such as Shang Yefeng proposed the method that Geodesic Main driving wheel exterior feature based on region shape priori is cut apart ultrasonoscopy cardiac valve; Abroad, 2006, the people such as Sebastien Martin proposed a kind ofly based on active contour model, to cut apart tricuspid semi-automatic method in ultrasonoscopy.But existing these methods, main is all for the long axial images in echocardiogram, while using these methods to cut apart aorta petal ultrasonoscopy, due to ultrasonoscopy edge fog and there is much noise, weak edge exists overflows in a large number, thereby makes segmentation result accurate not, credible.The problem existing based on existing method, researcher attempts the aorta petal of ultrasonic short axis images to cut apart, but the current research about ultrasonic short axis images dividing method is also seldom.

Summary of the invention

The object of this invention is to provide a kind of GCV model based on interframe shape constraining and cut apart aorta petal ultrasonic image sequence method, with solve that prior art exists for ultrasound long axis image, carry out aorta petal Ultrasound Image Segmentation time there is problem imperfect and that seriously overflow.

The object of the invention is to realize by following technical scheme:

GCV model based on interframe shape constraining is cut apart an aorta petal ultrasonic image sequence method, and it comprises the following steps:

A) obtain one group of continuous aorta petal ultrasonic image sequence, quantity is M, and extracts the sector region of each two field picture, and the threshold value of non-sector region is 255; Then each two field picture is carried out to Wiener filtering;

B) carry out, after Wiener filtering, calculating the gradient vector flow of each two field picture; And on each image, all define at random an initial evolution curve;

By calculating initial evolution curve method vector direction described in each and the cosine value of gradient vector flow angular separation described in each, gradient vector flow is joined to CV model framework as energy constraint item, obtain the GCV model of each image;

C) four points of manual definition on the 1st two field picture, then utilize B spline interpolation to form closed curve, using this closed curve as initial constraint shapes; Then utilize shape matching function to join in the GCV model of the 1st two field picture this initial constraint shapes as energy constraint item, then minimize the energy functional of this GCV model, obtain the aorta petal segmentation result of the 1st two field picture;

D) the aorta petal segmentation result of m two field picture is carried out to rolling ball method filtering, the result obtaining is as the non-initial constraint shapes of m+1 two field picture, then utilize shape matching function that this non-initial constraint shapes is joined in the GCV model of m+1 two field picture as energy constraint item, then minimize the energy functional of this GCV model, obtain the aorta petal segmentation result of m+1 two field picture; Wherein, 1≤m≤M-1.

Method of the present invention, b) described in step, calculate each two field picture gradient vector flow method specifically: minimization of energy function $E_{\mathrm{gvf}}\left(g\right)=\∫\∫\mathrm{\μ}(u_x^2+u_y^2+v_x^2+v_y^2)+{|\▿f|}^2{|g-\▿f|}^2\mathrm{dxdy},$ Obtain gradient vector flow g (x, y)=(u (x, y), v (x, y)); Wherein μ is regularization parameter, and f (x, y) represents the gradient map of ultrasonoscopy.

Method of the present invention, thus b) using gradient vector flow as energy constraint item, join formula that CV model framework obtains GCV model described in step specifically:

E ^gcv(φ,c ₁,c ₂)=E ^cv(φ,c ₁,c ₂)+αcos<n(φ),g>∫ _ΩH(φ)dxdy；

Wherein, $\begin{array}{c}{E}^{\mathrm{cv}}(\mathrm{\&phi;},c_1{,c}_2)=\mathrm{\&mu;}{\&Integral;}_{\mathrm{\&Omega;}}\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)|\&dtri;\mathrm{\&phi;}|\mathrm{dxdy}+v{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}\\ +{\mathrm{\&lambda;}}_1{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_1|}^{2}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+{\mathrm{\&lambda;}}_2{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_2|}^2(1-H\left(\mathrm{\&phi;}\right))\mathrm{dxdy}\end{array},$

The span of α is that 0.3～1, φ represents the curve that develops, n represent the to develop normal vector direction of curve φ, and g represents the gradient vector flow path direction of ultrasonoscopy,

get v=0, λ ₁=λ ₂=1.

Method of the present invention, d) described in step, rolling ball method filtering is, by spherical junctions constitutive element in mathematical morphology, the segmentation result of the aorta petal of m frame is carried out to burn into dilation operation, that is: R (F, B)=F o B=(F Θ B) ⊕ B; In formula, F represents the segmentation result of m frame, and B represents spherical junctions constitutive element, and its radius is got 15～22 pixels.

Method of the present invention, c) described in step by initial constraint shapes or d) formula that non-initial constraint shapes joined to GCV model as energy constraint item described in step is as follows:

$\begin{array}{c}E(\mathrm{\&phi;},c_1,c_2)=E^{\mathrm{gcv}}(\mathrm{\&phi;},c_1,c_2)+\mathrm{\&beta;}{E}_{\mathrm{shape}}\left(\mathrm{\&phi;}\right)\\ =\mathrm{\&mu;}{\&Integral;}_{\mathrm{\&Omega;}}\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)|\&dtri;\mathrm{\&phi;}|\mathrm{dxdy}+v{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}\\ +{\mathrm{\&lambda;}}_1{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_1|}^{2}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+{\mathrm{\&lambda;}}_2{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_2|}^2(1-H\left(\mathrm{\&phi;}\right))\mathrm{dxdy}\\ +\mathrm{\&alpha;}\cos <n\left(\mathrm{\&phi;}\right),g>{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+\mathrm{\&beta;}{\&Integral;}_{\mathrm{\&Omega;}}{(\mathrm{\&phi;}-{\mathrm{\&phi;}}_B)}^2\mathrm{dxdy}\end{array};$

In formula, E _shape(φ) represent shape energy constraint item, φ _brepresent c) initial constraint shapes or the d of step) the non-initial constraint shapes of step, the value of β is 0.05～0.2;

Then using Euler-Lagrange equation to minimize the above-mentioned energy functional about φ obtains:

$\frac{\&PartialD;\mathrm{\&phi;}}{\&PartialD;t}=\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)[\mathrm{\&mu;}\mathrm{div}\left(\frac{\&dtri;\mathrm{\&phi;}}{|\&dtri;\mathrm{\&phi;}|}\right)-{\mathrm{\&lambda;}}_1{(u-c_1)}^2+{\mathrm{\&lambda;}}_2{(u-c_2)}^2+\mathrm{\&alpha;}\frac{\&dtri;\mathrm{\&phi;}\&CenterDot;g}{|\&dtri;\mathrm{\&phi;}|\left|g\right|}+2\mathrm{\&beta;}(\mathrm{\&phi;}-{\mathrm{\&phi;}}_B)]=0,$ Its optimum solution is c) aorta petal segmentation result or the d of the 1st two field picture described in step) the aorta petal segmentation result of m+1 two field picture described in step.

The present invention be directed to that echocardiographic short axis images operates, by adding energy constraint item based on gradient vector flow and the energy constraint item based on interframe constraint shapes on the basis at CV model framework, overcome the impact of valve motion in cutting procedure, solved the problem that in aorta petal Ultrasound Image Segmentation, weak edge overflows; With CV model, compare, gradient vector flow add the weak marginal information of greatly having strengthened aorta petal ultrasonoscopy, can effectively suppress the problem that overflow at fuzzy edge place; Interframe shape constraining has not only reduced man-machine interactively, and can instruct evolution curve to approach objective contour; Meanwhile, adding of gradient vector flow and interframe shape constraining item, has reduced the restriction to evolution curve initial position, can define at random initial evolution curve.

The inventive method only needs four points of manual definition, just can from echocardiogram, be partitioned into accurately aorta petal, not only greatly reduced doctor's workload, and solved the problem of seriously overflowing in aorta petal Ultrasound Image Segmentation in prior art, its segmentation result is with there being the manual segmentation result that experience doctor completes to compare, its aliasing error is only 4.83%, can simply and efficiently be partitioned into aorta petal.

Accompanying drawing explanation

Fig. 1 is the inventive method process flow diagram.

Fig. 2 (a) is in obtained ultrasonic image sequence.

Fig. 2 (b) is that Fig. 2 (a) carries out the image after Wiener filtering.

Fig. 3 is the aortic area of the first two field picture.

Fig. 4 is the gradient vector flow of Fig. 3.

Fig. 5 (a) is the initial constraint shapes defining on initial frame image.

Fig. 5 (b) is the non-initial constraint shapes defining on non-initial frame image.

Fig. 6 (a) is the result of manually cutting apart aorta petal ultrasonoscopy.

Fig. 6 (b) is used the inventive method to cut apart the result of aorta petal ultrasonoscopy.

Fig. 6 (c) is used CV model to cut apart the result of aorta petal ultrasonoscopy.

Embodiment

Below in conjunction with accompanying drawing and a concrete example, the present invention will be further described.

The present embodiment exists

dual-Core CPU E5800@3.20GHz, video card is NVIDIA GeForce GT430NVIDIA GeForce GT430, inside save as 2.00GB, operating system is to realize in the computing machine of Window XP, and whole dividing method adopts C++ and Matlab language compilation.

The flow process of this method is undertaken by step as shown in Figure 1:

(1) obtain one group of continuous aorta petal ultrasonic image sequence, and extract the sector region of each frame, as Fig. 2 (a), the threshold value of non-sector region is 255; Then the image each frame being obtained carries out Wiener filtering, and filtered result is as Fig. 2 (b).

Carry out, after Wiener filtering processing, both can removing the speckle noise in ultrasonoscopy, again can well preserving edge information.

(2) build GCV model:

(2.1) calculate the gradient vector flow of each two field picture (as shown in Figure 3) aortic area: by minimization of energy function

$E_{\mathrm{gvf}}\left(g\right)=\&Integral;\&Integral;\mathrm{\&mu;}(u_x^2+u_y^2+v_x^2+v_y^2)+{|\&dtri;f|}^2{|g-\&dtri;f|}^2\mathrm{dxdy},$

Gradient vector flow g (x, y)=(u (x, y), the v (x, y)) that obtains this image, its result is as Fig. 4, and in Fig. 4, the direction of arrow represents gradient vector flow path direction;

Wherein, μ is regularization parameter, and f (x, y) represents the gradient map of ultrasonoscopy.

(2.2) (the evolution curve on each two field picture can be identical on each two field picture, to define at random an initial evolution curve, also can be different), by calculating on this image initial evolution curve method vector direction with the cosine value of gradient vector flow angular separation θ, gradient vector flow is added to CV model framework as new energy constraint item, foundation is for the GCV model of each two field picture, that is:

E ^gcv(φ,c ₁,c ₂)=E ^cv(φ,c ₁,c ₂)+αcos<n(φ),g>∫ _ΩH(φ)dxdy；

Wherein, $\begin{array}{c}{E}^{\mathrm{cv}}(\mathrm{\&phi;},c_1{,c}_2)=\mathrm{\&mu;}{\&Integral;}_{\mathrm{\&Omega;}}\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)|\&dtri;\mathrm{\&phi;}|\mathrm{dxdy}+v{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}\\ +{\mathrm{\&lambda;}}_1{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_1|}^{2}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+{\mathrm{\&lambda;}}_2{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_2|}^2(1-H\left(\mathrm{\&phi;}\right))\mathrm{dxdy}\end{array},$

The span of α is 0.3～1, the size for regulating gradient vector current field energy bound term on the impact of evolution curve; φ represents the curve that develops, n represent the to develop normal vector direction of curve φ, and g represents the gradient vector flow field direction of ultrasonoscopy,

for simplified operation, conventionally get v=0, λ ₁=λ ₂=1.

(4) four points of the aortic area manual definition on the first two field picture, then utilize B spline interpolation to form closed curve, are initial constraint shapes, as shown in Fig. 5 (a), this initial constraint shapes are joined in GCV model, that is: as energy constraint item

$\begin{array}{c}E(\mathrm{\&phi;},c_1,c_2)=E^{\mathrm{gcv}}(\mathrm{\&phi;},c_1,c_2)+\mathrm{\&beta;}{E}_{\mathrm{shape}}\left(\mathrm{\&phi;}\right)\\ =\mathrm{\&mu;}{\&Integral;}_{\mathrm{\&Omega;}}\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)|\&dtri;\mathrm{\&phi;}|\mathrm{dxdy}+v{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}\\ +{\mathrm{\&lambda;}}_1{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_1|}^{2}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+{\mathrm{\&lambda;}}_2{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_2|}^2(1-H\left(\mathrm{\&phi;}\right))\mathrm{dxdy}\\ +\mathrm{\&alpha;}\cos <n\left(\mathrm{\&phi;}\right),g>{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+\mathrm{\&beta;}{\&Integral;}_{\mathrm{\&Omega;}}{(\mathrm{\&phi;}-{\mathrm{\&phi;}}_B)}^2\mathrm{dxdy}\end{array},$

Wherein, E _shape(φ) represent shape energy constraint item, φ _brepresent initial constraint shapes, the span of β is 0.05～0.2, the size for adjusting energy bound term on the impact of evolution curve;

Then with Euler-Lagrange equation, minimizing the above-mentioned energy functional about φ obtains:

$\frac{\&PartialD;\mathrm{\&phi;}}{\&PartialD;t}=\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)[\mathrm{\&mu;}\mathrm{div}\left(\frac{\&dtri;\mathrm{\&phi;}}{|\&dtri;\mathrm{\&phi;}|}\right)-{\mathrm{\&lambda;}}_1{(u-c_1)}^2+{\mathrm{\&lambda;}}_2{(u-c_2)}^2+\mathrm{\&alpha;}\frac{\&dtri;\mathrm{\&phi;}\&CenterDot;g}{|\&dtri;\mathrm{\&phi;}|\left|g\right|}+2\mathrm{\&beta;}(\mathrm{\&phi;}-{\mathrm{\&phi;}}_B)]=0,$

Its optimum solution is the aorta petal segmentation result of the first two field picture.

(5) for the second two field picture, the aorta petal segmentation result of the first two field picture is carried out to the filtered result of rolling ball method as the non-initial constraint shapes of this present frame (i.e. the second frame), then this non-initial constraint shapes is joined in GCV model, that is: as energy constraint item

$\begin{array}{c}E(\mathrm{\&phi;},c_1,c_2)=E^{\mathrm{gcv}}(\mathrm{\&phi;},c_1,c_2)+\mathrm{\&beta;}{E}_{\mathrm{shape}}\left(\mathrm{\&phi;}\right)\\ =\mathrm{\&mu;}{\&Integral;}_{\mathrm{\&Omega;}}\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)|\&dtri;\mathrm{\&phi;}|\mathrm{dxdy}+v{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}\\ +{\mathrm{\&lambda;}}_1{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_1|}^{2}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+{\mathrm{\&lambda;}}_2{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_2|}^2(1-H\left(\mathrm{\&phi;}\right))\mathrm{dxdy}\\ +\mathrm{\&alpha;}\cos <n\left(\mathrm{\&phi;}\right),g>{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+\mathrm{\&beta;}{\&Integral;}_{\mathrm{\&Omega;}}{(\mathrm{\&phi;}-{\mathrm{\&phi;}}_B)}^2\mathrm{dxdy}\end{array},$

Wherein, E _shape(φ) represent shape constraining energy term, φ _brepresent constraint shapes, the span of β is 0.05～0.2, the size for adjustable shape bound energy type on the impact of evolution curve;

Then with Euler-Lagrange equation, minimizing the above-mentioned energy functional about φ obtains:

$\frac{\&PartialD;\mathrm{\&phi;}}{\&PartialD;t}=\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)[\mathrm{\&mu;}\mathrm{div}\left(\frac{\&dtri;\mathrm{\&phi;}}{|\&dtri;\mathrm{\&phi;}|}\right)-{\mathrm{\&lambda;}}_1{(u-c_1)}^2+{\mathrm{\&lambda;}}_2{(u-c_2)}^2+\mathrm{\&alpha;}\frac{\&dtri;\mathrm{\&phi;}\&CenterDot;g}{|\&dtri;\mathrm{\&phi;}|\left|g\right|}+2\mathrm{\&beta;}(\mathrm{\&phi;}-{\mathrm{\&phi;}}_B)]=0,$

Its optimum solution is the aorta petal segmentation result of the second two field picture.

For all images after the first frame, all according to the method for step (5), carry out, the aorta petal segmentation result (as shown in curve I in Fig. 5 (b)) that is about to previous frame image carries out the filtered result of rolling ball method as the non-initial constraint shapes (as shown in curve II in Fig. 5 (b)) of this present frame, then by the above-mentioned formula providing, calculate evolution, obtain the aorta petal segmentation result of this present frame.

It is, by spherical junctions constitutive element in mathematical morphology, the aorta petal segmentation result of previous frame is carried out to burn into dilation operation that above-mentioned spin is sent out filtering, that is: R (F, B)=F o B=(F Θ B) ⊕ B, in formula, F represents the segmentation result of consecutive frame, B represents spherical junctions constitutive element, and generally its radius is got 15～22 pixels.

For further verifying the feasibility of this method and the accuracy of its segmentation result, the aorta petal ultrasonoscopy of continuous 5 frames has been carried out respectively manually cutting apart and using CV model to cut apart, and this segmentation result of two kinds and the result that adopts the inventive method to cut apart are compared, result is as shown in Figure 6: Fig. 6 (a) is manually cut apart the result of aorta petal for experienced Ultrasonography doctor, Fig. 6 (b) is cut apart the result of aorta petal for the inventive method, and Fig. 6 (c) is for being used CV model to cut apart the result of aorta petal.From figure, obviously can find out, the result of using segmentation result that the inventive method obtains manually to cut apart with Ultrasonography doctor is about the same, and use CV model to carry out ultrasonoscopy aorta petal, cuts apart Ze Ruo edge and exists and overflow in a large number.

From above embodiment test findings, the invention solves the problem of seriously overflowing in aorta petal Ultrasound Image Segmentation in prior art, can greatly reduce doctor's workload, be partitioned into simply efficiently aorta petal.

Claims (5)

1. the GCV model based on interframe shape constraining is cut apart an aorta petal ultrasonic image sequence method, it is characterized in that comprising the following steps:

A) obtain one group of continuous aorta petal ultrasonic image sequence, quantity is M, and extracts the sector region of each two field picture, and the threshold value of non-sector region is 255; Then each two field picture is carried out to Wiener filtering;

B) carry out, after Wiener filtering, calculating the gradient vector flow of each two field picture, and on each image, all define at random an initial evolution curve;

By calculating initial evolution curve method vector direction described in each and the cosine value of gradient vector flow angular separation described in each, gradient vector flow is joined to CV model framework as energy constraint item, obtain the GCV model of each image;

C) four points of manual definition on the 1st two field picture, then utilize B spline interpolation to form closed curve, using this closed curve as initial constraint shapes; Then utilize shape matching function to join in the GCV model of the 1st two field picture this initial constraint shapes as energy constraint item, then minimize the energy functional of this GCV model, obtain the aorta petal segmentation result of the 1st two field picture;

D) the aorta petal segmentation result of m two field picture is carried out to rolling ball method filtering, the result obtaining is as the non-initial constraint shapes of m+1 two field picture, then utilize shape matching function that this non-initial constraint shapes is joined in the GCV model of m+1 two field picture as energy constraint item, then minimize the energy functional of this GCV model, obtain the aorta petal segmentation result of m+1 two field picture; Wherein, 1≤m≤M-1.

2. the GCV model based on interframe shape constraining according to claim 1 is cut apart aorta petal ultrasonic image sequence method, it is characterized in that, b) described in step, calculate each two field picture gradient vector flow method specifically: minimization of energy function $E_{\mathrm{gvf}}\left(g\right)=\&Integral;\&Integral;\mathrm{\&mu;}(u_x^2+u_y^2+v_x^2+v_y^2)+{|\&dtri;f|}^2{|g-\&dtri;f|}^2\mathrm{dxdy},$ Obtain gradient vector flow g (x, y)=(u (x, y), v (x, y)); Wherein μ is regularization parameter, and f (x, y) represents the gradient map of ultrasonoscopy.

3. the GCV model based on interframe shape constraining according to claim 1 is cut apart aorta petal ultrasonic image sequence method, it is characterized in that, thereby b) using gradient vector flow as energy constraint item, join formula that CV model framework obtains GCV model described in step specifically:

E ^gcv(φ,c ₁,c ₂)=E ^cv(φ,c ₁,c ₂)+αcos<n(φ),g>∫ _ΩH(φ)dxdy；

Wherein, $\begin{array}{c}{E}^{\mathrm{cv}}(\mathrm{\&phi;},c_1{,c}_2)=\mathrm{\&mu;}{\&Integral;}_{\mathrm{\&Omega;}}\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)|\&dtri;\mathrm{\&phi;}|\mathrm{dxdy}+v{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}\\ +{\mathrm{\&lambda;}}_1{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_1|}^{2}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+{\mathrm{\&lambda;}}_2{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_2|}^2(1-H\left(\mathrm{\&phi;}\right))\mathrm{dxdy}\end{array},$ The span of α is that 0.3～1, φ represents the curve that develops, n represent the to develop normal vector direction of curve φ, and g represents the gradient vector flow path direction of ultrasonoscopy, $\cos \mathrm{\&theta;}=\cos <n,g>=\frac{n\&CenterDot;g}{\left|n\right|\left|g\right|},$ Get v=0, λ ₁=λ ₂=1.

4. the GCV model based on interframe shape constraining according to claim 1 is cut apart aorta petal ultrasonic image sequence method, it is characterized in that, d) described in step, rolling ball method filtering is, by spherical junctions constitutive element in mathematical morphology, the segmentation result of the aorta petal of m frame is carried out to burn into dilation operation, that is: R (F, B)=F o B=(F Θ B) ⊕ B; In formula, F represents the segmentation result of m frame, and B represents spherical junctions constitutive element, and its radius is got 15～22 pixels.

5. the GCV model based on interframe shape constraining according to claim 1 is cut apart aorta petal ultrasonic image sequence method, it is characterized in that, c) described in step by initial constraint shapes or d) formula that non-initial constraint shapes joined to GCV model as energy constraint item described in step is as follows:

$\begin{array}{c}E(\mathrm{\&phi;},c_1,c_2)=E^{\mathrm{gcv}}(\mathrm{\&phi;},c_1,c_2)+\mathrm{\&beta;}{E}_{\mathrm{shape}}\left(\mathrm{\&phi;}\right)\\ =\mathrm{\&mu;}{\&Integral;}_{\mathrm{\&Omega;}}\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)|\&dtri;\mathrm{\&phi;}|\mathrm{dxdy}+v{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}\\ +{\mathrm{\&lambda;}}_1{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_1|}^{2}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+{\mathrm{\&lambda;}}_2{\&Integral;}_{\mathrm{\&Omega;}}{|u(x,y)-c_2|}^2(1-H\left(\mathrm{\&phi;}\right))\mathrm{dxdy}\\ +\mathrm{\&alpha;}\cos <n\left(\mathrm{\&phi;}\right),g>{\&Integral;}_{\mathrm{\&Omega;}}H\left(\mathrm{\&phi;}\right)\mathrm{dxdy}+\mathrm{\&beta;}{\&Integral;}_{\mathrm{\&Omega;}}{(\mathrm{\&phi;}-{\mathrm{\&phi;}}_B)}^2\mathrm{dxdy}\end{array};$

In formula, E _shape(φ) represent shape energy constraint item, φ _brepresent c) initial constraint shapes or the d of step) the non-initial constraint shapes of step, the value of β is 0.05～0.2;

Then using Euler-Lagrange equation to minimize the above-mentioned energy functional about φ obtains:

$\frac{\&PartialD;\mathrm{\&phi;}}{\&PartialD;t}=\mathrm{\&delta;}\left(\mathrm{\&phi;}\right)[\mathrm{\&mu;}\mathrm{div}\left(\frac{\&dtri;\mathrm{\&phi;}}{|\&dtri;\mathrm{\&phi;}|}\right)-{\mathrm{\&lambda;}}_1{(u-c_1)}^2+{\mathrm{\&lambda;}}_2{(u-c_2)}^2+\mathrm{\&alpha;}\frac{\&dtri;\mathrm{\&phi;}\&CenterDot;g}{|\&dtri;\mathrm{\&phi;}|\left|g\right|}+2\mathrm{\&beta;}(\mathrm{\&phi;}-{\mathrm{\&phi;}}_B)]=0,$ Its optimum solution is c) aorta petal segmentation result or the d of the 1st two field picture described in step) the aorta petal segmentation result of m+1 two field picture described in step.

Images