CN103886599B - A kind of blood vessel ROI dividing methods based on ivus image - Google Patents
A kind of blood vessel ROI dividing methods based on ivus image Download PDFInfo
- Publication number
- CN103886599B CN103886599B CN201410115141.8A CN201410115141A CN103886599B CN 103886599 B CN103886599 B CN 103886599B CN 201410115141 A CN201410115141 A CN 201410115141A CN 103886599 B CN103886599 B CN 103886599B
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- msup
- msubsup
- image
- 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.)
- Expired - Fee Related
Links
Landscapes
- Medicines Containing Antibodies Or Antigens For Use As Internal Diagnostic Agents (AREA)
- Image Analysis (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The present invention relates to a kind of blood vessel ROI dividing methods based on ivus image.Methods described splits the lumen area and luminal membrane profile of blood vessel first, by positioning the center of lumen area to obtain the initial profile of parameter active contour model, then the middle outer membrane contour curve of blood vessel is obtained by convergence, the extraction of middle outer membrane takes full advantage of the priori of lumen area information.Finally, by middle outer membrane contour curve using inner region as ROI, the segmentation to vascular plaque is realized by global minimization's movable contour model.The present invention realizes the profile information visualization of outer membrane and patch in blood vessel ROI luminal membrane, compared with based on statistical IVUS image partition methods, its complicated statistical modeling process is abandoned and segmentation result is not influenceed by IVUS image artifacts and Patch properties;The pre-segmentation step that initial profile is carried out to IVUS images China and foreign countries film edge is eliminated, segmentation efficiency is improved.
Description
Technical field
It is more particularly to a kind of to be based on intravascular ultrasound IVUS the present invention relates to computer medical image analysis field
Vessel region of interest ROI (region of interest) segmentation side of (Intravascular Ultrasound) image
Method.
Background technology
At present in most countries, cardiovascular and cerebrovascular disease progressively turns into one of main factor of human death.It is coronal
Atherosclerotic lesion is the Etiological for causing miocardial infarction and cerebral infarction, can be accomplished to hardening illness if in early stage
Become the identification and diagnosis of portion forms, this there are the diagnosis and treatment to coronary artery disease great meaning.IVUS is so
A kind of Ultrasonic Diagnosis method of angiocardiopathy, ivus image can show real-time vascular wall form patch for doctor
Form and then directive significance is provided to the clinical diagnosis of angiocardiopathy.
However, for gathering the IVUS images come, although its imaging can show the structure of vascular wall and the shape of patch
State, but in actual diagnosis and treatment, region (ROI) interested to doctor is vessel lumen region in IVUS images, China and foreign countries' diaphragm area, with
And the patch region between tube chamber and middle outer membrane.In the area-of-interest of IVUS images, doctor generally requires range estimation or root
According to the edge of outer membrane and patch profile in empirically determined luminal membrane, so just inevitably leading to diagnostic result can not be very
Objectively reflect actual conditions, while can also bring certain operational difficulties to doctor.
Therefore, quick and precisely automatic Ground Split is carried out to ivus image by computer and just seems necessary,
At present, the computer partitioning algorithm of ivus image mainly has three kinds:The first is based on statistical image segmentation side
Method, document 1 (G.Mendizabal-Ruiz, M.Rivera, et al., " A probabilistic segmentation
method for the identification of luminal borders in intrvascular ultrasound
images”,IEEE Conference on Computer Vision and pattern Recognition,pp.1-8,
2008.) statistics modeling is carried out by the intensity profile to IVUS images and realizes the segmentation of ivus image, but be due to
The presence of the complex characteristics of image such as artifact and patch in ivus image can substantially reduce statistical modeling
Accuracy.Second method mainly realizes that ivus image is split by the means of machine learning, but this method
Model is excessively complicated, is often limited in actual utilization by more, the third is the calculation based on movable contour model
Method, (Xinsun Feng Rong etc. is prolonged in palace to document 2, " the ivus image side based on blood speckles noise suppressed and T-Snake models
Edge is extracted ", Chinese image graphics journal, Vol.12, No.4, pp.655-660,2007) although the method proposed can obtain good
Good segmentation result, but dependent on the selection of initial profile line, it is higher particularly with edge relatively fuzzy, texture self-similarity
IVUS images, its initial profile is less susceptible to determine, so as to also have impact on the accuracy of this dividing method to a certain extent.
The content of the invention
For the above-mentioned problems in the prior art, the present invention provides a kind of blood vessel based on ivus image
ROI dividing methods, the lumen area and tube chamber of blood vessel are split using improved Level Set Models algorithm combination arrowband method first
Film profile, by positioning the center of lumen area to obtain parameter active contour model (Snake) initial profile, then passes through
Convergence obtains the middle outer membrane contour curve of blood vessel, and the extraction of middle outer membrane takes full advantage of the priori of lumen area information.Most
Afterwards, by middle outer membrane contour curve using inner region as ROI, realized by global minimization's movable contour model to vascular plaque
Segmentation.
Blood vessel ROI dividing methods based on ivus image, are based on ivus image, according to IVUS
The characteristics of image, carry out outer membrane in vessel lumen film segmentation, blood vessel successively to ivus image and split, vascular plaque point
Cut.This method can obtain lumen area profile information, and China and foreign countries' film edge information and patch shape profile information etc. are not increasing
On the premise of optional equipment, make full use of ultrasonoscopy half-tone information in itself that intravascular ultrasound equipment is provided and region special
Property and combine various partitioning algorithms, realize blood vessel ROI segmentation.
The feature of the present invention is as follows:
Step 1, using intravascular ultrasound instrument, conduit of at the uniform velocity pulling back obtains the intravascular ultrasound video of human coronaries
Image.
Step 2, intravascular ultrasound video image step 1 obtained imports computer, and continuous blood is intercepted from video
Intraductal ultrasonography image sequence is as experimental image, and image resolution ratio is 2x0*2y0, hereinafter referred to as ultrasonoscopy;
Step 3, the segmentation of ultrasonoscopy luminal membrane is carried out to above-mentioned ultrasonoscopy.
Step 3.1, create and the initial tube chamber contour mould matrix of IVUS images size identical to be split.
If image to be split is I (x, y), creates size and be similarly 2x0*2y0Empty matrix Minit, establishment centre coordinate is O
(x0,y0), make Minit(x0-a:x0+b,y0-c:y0+ d)=1, that is, create a width of a+b of a length of c+d rectangle initialization area mould
Plate.
Step 3.2, symbolic measurement mapping matrix is initialized.
If MinitFor the initial tube chamber contour mould matrix set in step 3.1, M is calculatedinitMiddle current pixel point with recently
Non-zero pixels point Euclidean distance, unit pixel distance be 1, if symbolic measurement mapping matrix is:
φ=bwdist (Minit)-bwdist(1-Minit)-0.5
Wherein, bwdist is range conversion function, and acquiescence calculates current pixel point and nearest non-zero pixel in binary map
Distance, and return to the matrix of consequence with former binary map formed objects.
Step 3.3, the width of arrowband is set.
Obtain after symbolic measurement mapping matrix φ, arrowband W is setnarrowbandWidth be 2k, arrowband Wnarrowband
The set constituted for the pixel of satisfaction-k≤φ≤k conditions.
Step 3.4, the view data power F in narrowband region is calculatedimage, formula is as follows:
Fimage=[IWnarrowband(x,y)-ca]2-[IWnarrowband(x,y)-cb]2
Wherein ca、cbIt is the gray average inside and outside image I (x, y) to be split on two regions respectively.
Step 3.5, curvature speed term V is calculatedcurve。
For the point φ (x, y) (- k≤φ≤k) of narrowband region, φ is constructed by the method for eight neighborhood respectivelyx、φy、
φxx、φyyAnd φxy:
φx=φ (x+1, y)-φ (x-1, y)
φy=φ (x, y+1)-φ (x, y-1)
φxx=φ (x-1, y)+φ (x+1, y) -2 φ (x, y)
φyy=φ (x, y-1)+φ (x, y+1) -2 φ (x, y)
φxy=-0.25* φ (x-1, y-1) -0.25* φ (x+1, y+1)
+0.25*φ(x+1,y-1)+0.25*φ(x-1,y+1)
Curvature speed term VcurveFor:
Step 3.6, calculating speed function, formula is as follows:
Wherein, α is fairing weight coefficient.
Step 3.7, to arrowband WnarrowbandInternal symbol distance function mapping matrix φ is iterated, and iterative formula is:
In formula, n is iterations, Δ t1=0.45/max (F) is time step.
Step 3.8, by n iteration, final symbolic measurement mapping matrix φ iteration result is vessel lumen
The two values matrix in region, is denoted as lumen area contour images for finit, it is to obtain luminal membrane segmentation result to extract tube chamber film edge.
Step 4, the middle outer membrane segmentation of ultrasonoscopy is realized.
Step 4.1, the center of tube chamber contour area image is determined.
Calculate the vessel lumen region contour image f obtained in step 3initP+q rank squares mpq, formula is as follows:
Wherein, p and q are nonnegative integers.
The center C of tube chamber contour area imagecentre(xc,yc) coordinate be:
Wherein, m00Represent image finit0 rank square, m10And m01Represent image finitFirst moment.
Step 4.2, Snake initial profiles are initialized.
If Snake initial profile curves are C0(s), its coordinate points is (x (s), y (s)), orderS ∈ [0,1], θ ∈ [0,2 π].
Step 4.3, the equation for asking Snake contour curves to meet.
Define C (s)=(x (s), y (s)) energy:
Wherein,For image energy, I (x, y) is image to be split.
S derivations are obtained:
If spatial mesh size is h, the coordinate put on contour line is Xi=(xi,yi), by derivative difference approximation:
Above formula is turned into XiVector form system of linear equations:
Wherein, A is symmetrical five diagonal circular matrixes being made up of α, β, h, and (x, y) is the abscissa put on contour line and indulged
The vector of coordinate composition, fx(x,y),fy(x, y) is by EextIn XiPlace is to x, the vector of y partial derivative composition.
Step 4.4, the middle epicardium contours of iterative blood vessel are passed through.
Profile C (s) is regarded as to s and time t function C (s, t),Just
It is worth and is:C (s, 0)=C0(s), iterative result is:
Initial Snake curves are chosen as the first solution of equation, when solving convergence, i.e. CiOptimal song is obtained during (s, t) ≈ 0
Line, the curve is the middle epicardium contours of blood vessel.
Step 5, vascular plaque part is split using global minimization's active contour model algorithm.
The ultrasonoscopy China and foreign countries film edge that step 4 is obtained is as ultrasonoscopy ROI critical line, institute in selecting step 4
The ROI that region within the middle outer membrane contour line of segmentation is split as patch, utilizes global minimization's active contour model algorithm
Vascular plaque part is split.
Step 5.1, construction needs the functional minimized.
Based on global minimization's movable contour model thought, in order to obtain ivus image ROI patch profile, structure
Make the functional for needing to minimize:
E1(u=ψΩc,c1,c2, λ) and=∫cgds+λ∫Ω((c1-f(x))2-(c2-f(x))2ψΩcdx
=TVg(ψΩc)+λ∫Ωr1(x,c1,c2)ψΩcdx
Wherein, TVg(u) be with weighting function g (x) function u total variation function, g (x) is edge indicator function,
Function u is characteristic function ψΩc。
Step 5.2, structure realm item of information, expression formula is:
r1(x,c1,c2)=(fROI(x)-c1)2-(fROI(x)-c2)2
Wherein, c1For the gray average in the region of u in image I (x) >=0.5, c2For the gray scale in the regions of u < 0.5 in image I (x)
Average, fROI(x) it is intravascular ultrasound ROI image to be split.
Step 5.3, regularizing functionals E1(u=ψΩc,c1,c2, λ):
Step 5.4, fixed v, searches for u conductsSolution:
U=v- θ divp
Wherein, p=(p1,p2), solved equation by fixed point methodObtain:
p0=0, and
Wherein,
Step 5.5, fixed u, searches for v conductsSolution:
V=min { max { u (x)-θ λ r1(x,c1,c2),0},1}
Step 5.6, patch is completed by iteration to split.
Initialize u0=0, p0=0, setting maximum iteration imax=N, is calculated respectively according to step 5.2,5.4 and 5.5
pi, ui,vi, stop iteration when iterations reaches the times N of setting, preserve current variable vNI.e. patch divides
Cut result.
Compared with prior art, the invention has the advantages that:
The present invention proposes a kind of ROI dividing methods based on ivus image, realizes blood vessel ROI luminal membrane
The profile information visualization of middle outer membrane and patch, compared with based on statistical IVUS image partition methods, the present invention is abandoned
Its complicated statistical modeling process and segmentation result is not influenceed by IVUS image artifacts and Patch properties;With it is traditional based on
The IVUS image partition methods of movable contour model are compared, and the present invention only need to be by setting simple rectangle initial profile template i.e.
Lumen area profile can be obtained successively to position Snake initial profiles with the segmentation of outer membrane in carrying out, and eliminate and IVUS is schemed
The step for carrying out the pre-segmentation of initial profile as China and foreign countries' film edge, improves segmentation efficiency.Invention introduces vascular plaque
Segmentation, this is that current IVUS image partition methods are not directed to.
Brief description of the drawings
Fig. 1 is IVUS image blood vessel ROI segmentation group figures:(a) it is IVUS images to be split, (b) is vessel lumen film point
Result is cut, (c) is the tube chamber contour area of binaryzation, and (d) is outer membrane segmentation result in blood vessel, and (e) is blood vessel ROI image, (f)
Shown for blood vessel ROI patch profile;
Fig. 2 is continuous 9 frame ivus image luminal membrane segmentation effect figure;
Fig. 3 is outer membrane segmentation effect figure in continuous 9 frame ivus image;
Fig. 4 is continuous 9 frame ivus image ROI patch profile diagrams;
Fig. 5 is the flow chart of the inventive method.
Embodiment
The present invention is realized using following technological means:
A kind of blood vessel ROI dividing methods based on ivus image.First, segmentation figure picture is treated with reference to improved water
Flat collection model algorithm and arrowband method realize the segmentation to ivus image luminal membrane, then, position the center of lumen area
And generate the initial profile curve of Snake models and be iterated convergence so as to obtain the edge of outer membrane in blood vessel, finally, in selection
ROI of the outer membrane using inner region as patch, is realized to blood vessel ROI patch wheel with reference to the algorithm of global minimization's active contour model
Wide segmentation.
The above-mentioned blood vessel ROI dividing methods based on ivus image, comprise the steps:
Step 1, using intravascular ultrasound instrument, at the uniform velocity pulled back conduit with 0.5mm/s speed, obtain human coronaries'
Intravascular ultrasound video image;
Step 2, intravascular ultrasound video image step (1) obtained imports computer, is intercepted from video continuous
Intravascular ultrasound image sequence is as experimental image, and image resolution ratio is 384*384, hereinafter referred to as ultrasonoscopy;
Step 3, the segmentation of vessel lumen film is carried out to ultrasonoscopy;
If image to be split is I (x, y), the empty matrix M that size is similarly 384*384 is createdinit, establishing centre coordinate is
O (192,192), makes Minit(192-a:192+b,192-c:192+d)=1, that is, at the beginning of the rectangle for creating a width of a+b of a length of c+d
Beginningization region template.
If MinitFor initial profile pattern matrix, M is calculatedinitMiddle current pixel point and the Euclidean of nearest non-zero pixels point
Distance (unit pixel distance is 1), if symbolic measurement mapping matrix:
φ=bwdist (Minit)-bwdist(1-Minit)-0.5 (1)
Wherein, bwdist is range conversion function, and acquiescence calculates current pixel point and nearest non-zero pixel in binary map
Distance, and return to the matrix of consequence with former binary map formed objects.
The width for setting arrowband is 2k, if arrowband WnarrowbandThe collection constituted for the pixel of satisfaction-k≤φ≤k conditions
Close.Set up an office collection PinnerFor the set for the pixel for meeting condition (φ≤0), point set PouterTo meet condition (φ > 0) pixel
The set of point, and image I (x, y) to be split is calculated in PinnerAnd PouterThe gray average in region.
Construct the view data power F in narrowband regionimage:
Fimage=[IWnarrowband(x,y)-ca]2-[IWnarrowband(x,y)-cb]2 (2)
For the point φ (x, y) (- k≤φ≤k) of narrowband region, φ is constructed using the method for eight neighborhood respectivelyx,φy,
φxx,φyy,φxy
φx=φ (x+1, y)-φ (x-1, y) (3)
φy=φ (x, y+1)-φ (x, y-1) (4)
φxx=φ (x-1, y)+φ (x+1, y) -2 φ (x, y) (5)
φyyThe φ (x, y) (6) of=φ (x, y-1)+φ (x, y+1) -2
φxy=-0.25* φ (x-1, y-1) -0.25* φ (x+1, y+1)+0.25* φ (x+1, y-1)+0.25* φ (x-1,
y+1) (7)
Calculated by following formula and obtain curvature speed term:
The level set velocity function F of computed improved:
Wherein fairing weight coefficient α is constant, passes through arrowband WnarrowbandInterior symbolic measurement iterative formulaThe iteration of symbolic measurement is carried out, wherein n is iterations, and time step is set to Δ t1=
0.45/max(F).By n iteration, final iteration result φ is the two values matrix in vessel lumen region, extracts luminal membrane side
Edge, it is f to remember the two values matrixinit, such as shown in Fig. 1 (c).
Step 4, the segmentation of outer membrane in ultrasonoscopy is realized;
For the vessel lumen region contour image f obtained in step 3init, it is assumed that:
mpqFor image finitP+q rank squares, wherein p and q are nonnegative integer, the center C of vessel lumencentre(xc,yc)
Coordinate be calculated as follows:
Snake initial profile curves C is obtained by following formula0(s) coordinate (x of the point onsnake,ysnake):
Wherein, θ ∈ [0,2 π].
OrderWhereins∈[0,1].The energy for defining curve C (s)=(x (s), y (s)) is:
Image energy EextFor:
Wherein I (x, y) is image to be split.Above formula derivation is obtained:
If spatial mesh size is h, the coordinate put on contour line is Xi=(xi,yi), derivative is obtained with difference approximation:
Above formula can turn to following XiVector form system of linear equations:
Wherein A is symmetrical five diagonal circular matrixes being made up of α, β, h, and (x, y) is the abscissa put on contour line and indulged
The vector of coordinate composition, fx(x,y),fy(x, y) is by EextIn XiPlace is to x, the vector of y partial derivative composition.
Profile C is regarded as to s and time t function C (s, t), then:
Iterative result is:
Take initial value C (s, 0)=C0(s) as the first solution of equation, after n times iteration when solve convergence when be Ci(s,t)≈0
When obtain optimal curve, the curve is that the coordinate put on the middle epicardium contours of blood vessel, contour line is (xfin,yfin)。
Step 5, the ROI image of ultrasonoscopy is loaded into, is calculated shown in such as Fig. 1 (e) using global minimization's active contour model
Method, splits blood vessel ROI patch profiles.
The middle outer membrane contour curve obtained by step 4, outer membrane contour curve interior zone is blood vessel ROI in selection, is obtained
To the blood vessel ROI image f of ultrasonoscopyROI.To obtain ivus image ROI patch profile, lived based on global minimization
Dynamic skeleton pattern thought construction needs the functional minimized
Wherein TVg(u) it is that the function u total variation function g (x) with weighting function g (x) is edge indicator function, letter
Number u is characteristic function ψΩc。
Structure realm item of information:
r1(x,c1,c2)=(fROI(x)-c1)2-(fROI(x)-c2)2 (22)
Wherein c1For the area grayscale average of u in image I (x) >=0.5, c2For the area grayscale averages of u < in image I (x) 0.5.
fROI(x) it is blood vessel ROI image to be split.
Regularization E1(u=ψΩc,c1,c2, λ) and obtain following formula:
U is searched for by fixed v to be used asSolution, solve:
U=v- θ divp (24)
Wherein p=(p1,p2) be given by,
Above formula can be solved by fixed point method:
Wherein p0=0,
V is searched for by fixed u to be used asSolution, solve:
V=min { max { u (x)-θ λ r1(x,c1,c2),0},1} (27)
In the iterative process that patch is split, u is initialized0=0, p0=0, iterations i is set, according to formula (26) (24)
(22) (27) calculate p respectivelyi, ui,vi, when iterations reaches the times N of setting, stop iteration, preserve
Current variable vNThat is patch segmentation result, shown in such as Fig. 1 (f).
Claims (2)
1. a kind of blood vessel ROI dividing methods based on ivus image, it is characterised in that comprise the following steps:
Step 1, using intravascular ultrasound instrument, conduit of at the uniform velocity pulling back obtains the intravascular ultrasound video image of human coronaries;
Step 2, intravascular ultrasound video image step 1 obtained imports computer, is intercepted from video continuous intravascular
Ultrasonic image sequence is as experimental image, and image resolution ratio is 2x0*2y0, hereinafter referred to as ultrasonoscopy;
Step 3, the segmentation of ultrasonoscopy luminal membrane is carried out to above-mentioned ultrasonoscopy;
Step 3.1, create and the initial tube chamber contour mould matrix of ivus image size identical to be split;
If image to be split is I (x, y), creates size and be similarly 2x0*2y0Empty matrix Minit, establishment centre coordinate is O (x0,
y0), make Minit(x0-a:x0+b,y0-c:y0+ d)=1, that is, create a width of a+b of a length of c+d rectangle initialization area template;
Step 3.2, symbolic measurement mapping matrix is initialized;
If MinitFor the initial tube chamber contour mould matrix set in step 3.1, M is calculatedinitMiddle current pixel point and recently non-
The Euclidean distance of zero pixel, unit pixel distance is 1, if symbolic measurement mapping matrix is:
φ=bwdist (Minit)-bwdist(1-Minit)-0.5
Wherein, bwdist is range conversion function, acquiescence calculate in binary map current pixel point and nearest non-zero pixel away from
From, and return to the matrix of consequence with former binary map formed objects;
Step 3.3, the width of arrowband is set;
Obtain after symbolic measurement mapping matrix φ, arrowband W is setnarrowbandWidth be 2k, arrowband WnarrowbandIt is full
The set of the pixel composition of foot-k≤φ≤k conditions;
Step 3.4, the view data power F in narrowband region is calculatedimage, formula is as follows:
Fimage=[IWnarrowband(x,y)-ca]2-[IWnarrowband(x,y)-cb]2
Wherein ca、cbIt is the gray average inside and outside image I (x, y) to be split on two regions respectively;
Step 3.5, curvature speed term V is calculatedcurve;
For the point φ (x, y) of narrowband region ,-k≤φ≤k, φ is constructed by the method for eight neighborhood respectivelyx、φy、φxx、φyy
And φxy:
φx=φ (x+1, y)-φ (x-1, y)
φy=φ (x, y+1)-φ (x, y-1)
φxx=φ (x-1, y)+φ (x+1, y) -2 φ (x, y)
φyy=φ (x, y-1)+φ (x, y+1) -2 φ (x, y)
φxy=-0.25* φ (x-1, y-1) -0.25* φ (x+1, y+1)
+0.25*φ(x+1,y-1)+0.25*φ(x-1,y+1)
Curvature speed term VcurveFor:
<mrow>
<msub>
<mi>V</mi>
<mrow>
<mi>c</mi>
<mi>u</mi>
<mi>r</mi>
<mi>v</mi>
<mi>e</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>&phi;</mi>
<mrow>
<mi>x</mi>
<mi>x</mi>
</mrow>
</msub>
<msubsup>
<mi>&phi;</mi>
<mi>y</mi>
<mn>2</mn>
</msubsup>
<mo>-</mo>
<mn>2</mn>
<msub>
<mi>&phi;</mi>
<mi>y</mi>
</msub>
<msub>
<mi>&phi;</mi>
<mi>x</mi>
</msub>
<msub>
<mi>&phi;</mi>
<mrow>
<mi>x</mi>
<mi>y</mi>
</mrow>
</msub>
<mo>+</mo>
<msub>
<mi>&phi;</mi>
<mrow>
<mi>y</mi>
<mi>y</mi>
</mrow>
</msub>
<msubsup>
<mi>&phi;</mi>
<mi>x</mi>
<mn>2</mn>
</msubsup>
</mrow>
<msup>
<mrow>
<mo>(</mo>
<msubsup>
<mi>&phi;</mi>
<mi>x</mi>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<msubsup>
<mi>&phi;</mi>
<mi>y</mi>
<mn>2</mn>
</msubsup>
<mo>)</mo>
</mrow>
<mfrac>
<mn>3</mn>
<mn>2</mn>
</mfrac>
</msup>
</mfrac>
</mrow>
Step 3.6, calculating speed function, formula is as follows:
<mrow>
<mi>F</mi>
<mo>=</mo>
<mfrac>
<msub>
<mi>F</mi>
<mrow>
<mi>i</mi>
<mi>m</mi>
<mi>a</mi>
<mi>g</mi>
<mi>e</mi>
</mrow>
</msub>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
<mrow>
<mo>(</mo>
<mo>|</mo>
<msub>
<mi>F</mi>
<mrow>
<mi>i</mi>
<mi>m</mi>
<mi>a</mi>
<mi>g</mi>
<mi>e</mi>
</mrow>
</msub>
<mo>|</mo>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>+</mo>
<mi>&alpha;</mi>
<mo>&CenterDot;</mo>
<mfrac>
<mrow>
<msub>
<mi>&phi;</mi>
<mrow>
<mi>x</mi>
<mi>x</mi>
</mrow>
</msub>
<msubsup>
<mi>&phi;</mi>
<mi>y</mi>
<mn>2</mn>
</msubsup>
<mo>-</mo>
<mn>2</mn>
<msub>
<mi>&phi;</mi>
<mi>y</mi>
</msub>
<msub>
<mi>&phi;</mi>
<mi>x</mi>
</msub>
<msub>
<mi>&phi;</mi>
<mrow>
<mi>x</mi>
<mi>y</mi>
</mrow>
</msub>
<mo>+</mo>
<msub>
<mi>&phi;</mi>
<mrow>
<mi>y</mi>
<mi>y</mi>
</mrow>
</msub>
<msubsup>
<mi>&phi;</mi>
<mi>x</mi>
<mn>2</mn>
</msubsup>
</mrow>
<msup>
<mrow>
<mo>(</mo>
<msubsup>
<mi>&phi;</mi>
<mi>x</mi>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<msubsup>
<mi>&phi;</mi>
<mi>y</mi>
<mn>2</mn>
</msubsup>
<mo>)</mo>
</mrow>
<mfrac>
<mn>3</mn>
<mn>2</mn>
</mfrac>
</msup>
</mfrac>
</mrow>
Wherein, α is fairing weight coefficient;
Step 3.7, to arrowband WnarrowbandInternal symbol distance function mapping matrix φ is iterated, and iterative formula is:
<mrow>
<msubsup>
<mi>&phi;</mi>
<mi>W</mi>
<mrow>
<mi>n</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msubsup>
<mo>=</mo>
<msubsup>
<mi>&phi;</mi>
<mi>W</mi>
<mi>n</mi>
</msubsup>
<mo>+</mo>
<msub>
<mi>&Delta;t</mi>
<mn>1</mn>
</msub>
<mo>&CenterDot;</mo>
<mi>F</mi>
</mrow>
In formula, n is iterations, Δ t1=0.45/max (F) is time step;
Step 3.8, by n times iteration, N represents maximum iteration, final symbolic measurement mapping matrix φ iteration knot
Fruit is the two values matrix in vessel lumen region, is denoted as lumen area contour images for finit, it is to obtain to extract tube chamber film edge
Luminal membrane segmentation result;
Step 4, the middle outer membrane segmentation of ultrasonoscopy is realized;
Step 4.1, the center of tube chamber contour area image is determined;
Calculate the vessel lumen region contour image f obtained in step 3initP+q rank squares mpq, formula is as follows:
<mrow>
<msub>
<mi>m</mi>
<mrow>
<mi>p</mi>
<mi>q</mi>
</mrow>
</msub>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<msup>
<mi>i</mi>
<mi>p</mi>
</msup>
<msup>
<mi>j</mi>
<mi>q</mi>
</msup>
<msub>
<mi>f</mi>
<mrow>
<mi>i</mi>
<mi>n</mi>
<mi>i</mi>
<mi>t</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
Wherein, p and q are nonnegative integers;
The center C of tube chamber contour area imagecentre(xc,yc) coordinate be:
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>x</mi>
<mi>c</mi>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>m</mi>
<mn>10</mn>
</msub>
<msub>
<mi>m</mi>
<mn>00</mn>
</msub>
</mfrac>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>y</mi>
<mi>c</mi>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>m</mi>
<mn>01</mn>
</msub>
<msub>
<mi>m</mi>
<mn>00</mn>
</msub>
</mfrac>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
Wherein, m00Represent image finit0 rank square, m10And m01Represent image finitFirst moment;
Step 4.2, Snake initial profiles are initialized;
If Snake initial profile curves are C0(s), its coordinate points is (x (s), y (s)), orderS ∈ [0,1], θ ∈ [0,2 π];
Step 4.3, the equation for asking Snake contour curves to meet;
Define C (s)=(x (s), y (s)) energy:
<mrow>
<msub>
<mi>E</mi>
<mrow>
<mi>s</mi>
<mi>n</mi>
<mi>a</mi>
<mi>k</mi>
<mi>e</mi>
</mrow>
</msub>
<mo>=</mo>
<munderover>
<mo>&Integral;</mo>
<mn>0</mn>
<mn>1</mn>
</munderover>
<mo>{</mo>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<mo>&lsqb;</mo>
<mi>&alpha;</mi>
<mo>|</mo>
<msup>
<mi>C</mi>
<mo>&prime;</mo>
</msup>
<mrow>
<mo>(</mo>
<mi>s</mi>
<mo>)</mo>
</mrow>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>+</mo>
<mi>&beta;</mi>
<mo>|</mo>
<msup>
<mi>C</mi>
<mrow>
<mo>&prime;</mo>
<mo>&prime;</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>s</mi>
<mo>)</mo>
</mrow>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>&rsqb;</mo>
<mo>+</mo>
<msub>
<mi>E</mi>
<mrow>
<mi>e</mi>
<mi>x</mi>
<mi>t</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mi>C</mi>
<mo>(</mo>
<mi>s</mi>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mo>}</mo>
<mi>d</mi>
<mi>s</mi>
</mrow>
Wherein,For image energy, I (x, y) is image to be split;
S derivations are obtained:
<mrow>
<msup>
<mi>&alpha;C</mi>
<mrow>
<mo>&prime;</mo>
<mo>&prime;</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>s</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<msup>
<mi>&beta;C</mi>
<mrow>
<mo>&prime;</mo>
<mo>&prime;</mo>
<mo>&prime;</mo>
<mo>&prime;</mo>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>s</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>&dtri;</mo>
<msub>
<mi>E</mi>
<mrow>
<mi>e</mi>
<mi>x</mi>
<mi>t</mi>
</mrow>
</msub>
<mo>=</mo>
<mn>0</mn>
</mrow>
If spatial mesh size is h, the coordinate put on contour line is Xi=(xi,yi), by derivative difference approximation,
<mrow>
<msubsup>
<mi>X</mi>
<mi>i</mi>
<mrow>
<mo>&prime;</mo>
<mo>&prime;</mo>
</mrow>
</msubsup>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>X</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>-</mo>
<mn>2</mn>
<msub>
<mi>X</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msub>
<mi>X</mi>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
</mrow>
<msup>
<mi>h</mi>
<mn>2</mn>
</msup>
</mfrac>
<mo>,</mo>
<msubsup>
<mi>X</mi>
<mi>i</mi>
<mrow>
<mo>&prime;</mo>
<mo>&prime;</mo>
<mo>&prime;</mo>
<mo>&prime;</mo>
</mrow>
</msubsup>
<mo>=</mo>
<mfrac>
<mrow>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<msub>
<mi>X</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mn>2</mn>
<msub>
<mi>X</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>+</mo>
<msub>
<mi>X</mi>
<mi>i</mi>
</msub>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mo>-</mo>
<mn>2</mn>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<msub>
<mi>X</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>-</mo>
<mn>2</mn>
<msub>
<mi>X</mi>
<mi>i</mi>
</msub>
<mo>+</mo>
<msub>
<mi>X</mi>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mo>+</mo>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<msub>
<mi>X</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mn>2</mn>
<msub>
<mi>X</mi>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>+</mo>
<msub>
<mi>X</mi>
<mrow>
<mi>i</mi>
<mo>+</mo>
<mn>2</mn>
</mrow>
</msub>
</mrow>
<mo>&rsqb;</mo>
</mrow>
</mrow>
<msup>
<mi>h</mi>
<mn>4</mn>
</msup>
</mfrac>
</mrow>
Above formula is turned into XiVector form system of linear equations:
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mi>A</mi>
<mi>x</mi>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>x</mi>
</msub>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
<mo>=</mo>
<mn>0</mn>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>A</mi>
<mi>y</mi>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>y</mi>
</msub>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
<mo>=</mo>
<mn>0</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
Wherein, A is symmetrical five diagonal circular matrixes being made up of α, β, h, and (x, y) is the abscissa and ordinate put on contour line
The vector of composition, fx(x,y),fy(x, y) is by EextIn XiPlace is to x, the vector of y partial derivative composition;
Step 4.4, the middle epicardium contours of iterative blood vessel are passed through;
Profile C (s) is regarded as to s and time t function C (s, t),Initial value is:
C (s, 0)=C0(s), iterative result is:
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<msup>
<mrow>
<mo>(</mo>
<mi>A</mi>
<mo>+</mo>
<mi>&gamma;</mi>
<mi>I</mi>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>&gamma;x</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>f</mi>
<mi>x</mi>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>y</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>y</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<msup>
<mrow>
<mo>(</mo>
<mi>A</mi>
<mo>+</mo>
<mi>&gamma;</mi>
<mi>I</mi>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<msub>
<mi>&gamma;y</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>f</mi>
<mi>y</mi>
</msub>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>y</mi>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msub>
</mrow>
<mo>)</mo>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
Initial Snake curves are chosen as the first solution of equation, when solving convergence, i.e. CiOptimal curve is obtained during (s, t) ≈ 0, should
Curve is the middle epicardium contours of blood vessel;
Step 5, vascular plaque part is split using global minimization's active contour model algorithm;
The ultrasonoscopy China and foreign countries film edge that step 4 is obtained is split in selecting step 4 as ultrasonoscopy ROI critical line
Middle outer membrane contour line within the ROI split as patch of region, using global minimization's active contour model algorithm to blood
Pipe plaque components are split.
2. a kind of blood vessel ROI dividing methods based on ivus image according to claim 1, it is characterised in that
The method that the step 5 is split using global minimization's active contour model algorithm to vascular plaque part includes following step
Suddenly:
Step 5.1, construction needs the functional minimized;
Based on global minimization's movable contour model thought, in order to obtain ivus image ROI patch profile, construction is needed
The functional to be minimized:
<mfenced open = "" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>E</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<mi>u</mi>
<mo>=</mo>
<msub>
<mi>&psi;</mi>
<mrow>
<mi>&Omega;</mi>
<mi>c</mi>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>c</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>c</mi>
<mn>2</mn>
</msub>
<mo>,</mo>
<mi>&lambda;</mi>
</mrow>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mo>&Integral;</mo>
<mi>c</mi>
</msub>
<mi>g</mi>
<mi>d</mi>
<mi>s</mi>
<mo>+</mo>
<mi>&lambda;</mi>
<msub>
<mo>&Integral;</mo>
<mi>&Omega;</mi>
</msub>
<mo>(</mo>
<msup>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>c</mi>
<mn>1</mn>
</msub>
<mo>-</mo>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
</mrow>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>-</mo>
<msup>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>c</mi>
<mn>2</mn>
</msub>
<mo>-</mo>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
</mrow>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<msub>
<mi>&psi;</mi>
<mrow>
<mi>&Omega;</mi>
<mi>c</mi>
</mrow>
</msub>
<mi>d</mi>
<mi>x</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>=</mo>
<msub>
<mi>TV</mi>
<mi>g</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>&psi;</mi>
<mrow>
<mi>&Omega;</mi>
<mi>c</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>&lambda;</mi>
<msub>
<mo>&Integral;</mo>
<mi>&Omega;</mi>
</msub>
<msub>
<mi>r</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<mi>x</mi>
<mo>,</mo>
<msub>
<mi>c</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>c</mi>
<mn>2</mn>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<msub>
<mi>&psi;</mi>
<mrow>
<mi>&Omega;</mi>
<mi>c</mi>
</mrow>
</msub>
<mi>d</mi>
<mi>x</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
Wherein, TVg(u) be with weighting function g (x) function u total variation function, g (x) is edge indicator function, function u
It is characteristic function ψΩc;
Step 5.2, structure realm item of information, expression formula is:
r1(x,c1,c2)=(fROI(x)-c1)2-(fROI(x)-c2)2
Wherein, c1For the gray average in the region of u in image I (x) >=0.5, c2Gray scale for the regions of u < 0.5 in image I (x) is equal
Value, fROI(x) it is intravascular ultrasound ROI image to be split;
Step 5.3, regularizing functionals E1(u=ψΩc,c1,c2, λ):
<mrow>
<msubsup>
<mi>E</mi>
<mn>2</mn>
<mi>r</mi>
</msubsup>
<mrow>
<mo>(</mo>
<mi>u</mi>
<mo>,</mo>
<msub>
<mi>c</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>c</mi>
<mn>2</mn>
</msub>
<mo>,</mo>
<mi>&lambda;</mi>
<mo>,</mo>
<mi>&alpha;</mi>
<mo>,</mo>
<mi>&theta;</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mi>TV</mi>
<mi>g</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>u</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mn>2</mn>
<mi>&theta;</mi>
</mrow>
</mfrac>
<mo>|</mo>
<mo>|</mo>
<mi>u</mi>
<mo>-</mo>
<mi>v</mi>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<msup>
<mi>L</mi>
<mn>2</mn>
</msup>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<msub>
<mo>&Integral;</mo>
<mi>&Omega;</mi>
</msub>
<msub>
<mi>&lambda;r</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<msub>
<mi>c</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>c</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mi>v</mi>
<mo>+</mo>
<mi>&alpha;</mi>
<mi>w</mi>
<mrow>
<mo>(</mo>
<mi>v</mi>
<mo>)</mo>
</mrow>
<mi>d</mi>
<mi>x</mi>
</mrow>
Step 5.4, fixed v, searches for u conductsSolution:
U=v- θ divp
Wherein, p=(p1,p2), solved equation by fixed point methodObtain:
p0=0, and
Wherein,
Step 5.5, fixed u, searches for v conductsSolution:
V=min { max { u (x)-θ λ r1(x,c1,c2),0},1}
Step 5.6, patch is completed by iteration to split;
Initialize u0=0, p0=0, setting maximum iteration imax=N, p is calculated according to step 5.2,5.4 and 5.5 respectivelyi, ui,vi, stop iteration when iterations reaches the times N of setting, preserve current variable vNThat is patch segmentation result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410115141.8A CN103886599B (en) | 2014-03-26 | 2014-03-26 | A kind of blood vessel ROI dividing methods based on ivus image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410115141.8A CN103886599B (en) | 2014-03-26 | 2014-03-26 | A kind of blood vessel ROI dividing methods based on ivus image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103886599A CN103886599A (en) | 2014-06-25 |
CN103886599B true CN103886599B (en) | 2017-10-13 |
Family
ID=50955472
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410115141.8A Expired - Fee Related CN103886599B (en) | 2014-03-26 | 2014-03-26 | A kind of blood vessel ROI dividing methods based on ivus image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103886599B (en) |
Families Citing this family (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104537645A (en) * | 2014-12-15 | 2015-04-22 | 北京工业大学 | ROI mark point matching method based on intravascular unltrasound image |
CN106388867A (en) * | 2016-09-28 | 2017-02-15 | 深圳华声医疗技术有限公司 | Automatic identification measurement method for intima-media membrane in blood vessel and ultrasonic apparatus |
US11341633B2 (en) * | 2017-05-31 | 2022-05-24 | Edan Instruments, Inc. | Systems and methods for adaptive enhancement of vascular imaging |
CN107909581B (en) * | 2017-11-03 | 2019-01-29 | 杭州依图医疗技术有限公司 | Lobe of the lung section dividing method, device, system, storage medium and the equipment of CT images |
CN107909585B (en) * | 2017-11-14 | 2020-02-18 | 华南理工大学 | Intravascular intima segmentation method of intravascular ultrasonic image |
CN109166132B (en) * | 2018-07-16 | 2022-01-07 | 哈尔滨工程大学 | Side-scan sonar image target identification method with variable initial distance symbolic function |
CN109674493B (en) * | 2018-11-28 | 2021-08-03 | 深圳蓝韵医学影像有限公司 | Method, system and equipment for medical ultrasonic automatic tracking of carotid artery blood vessel |
CN110033466B (en) * | 2019-04-01 | 2020-12-18 | 数坤(北京)网络科技有限公司 | Coronary artery straightening image segmentation boundary determination method based on multiple gray levels |
CN110246136B (en) * | 2019-05-29 | 2021-07-02 | 山东大学 | Intravascular ultrasound parameter extraction method and system based on hybrid algorithm |
CN112258533B (en) * | 2020-10-26 | 2024-02-02 | 大连理工大学 | Method for segmenting cerebellum earthworm part in ultrasonic image |
CN112826535B (en) * | 2020-12-31 | 2022-09-09 | 青岛海信医疗设备股份有限公司 | Method, device and equipment for automatically positioning blood vessel in ultrasonic imaging |
CN113222956B (en) * | 2021-05-25 | 2023-09-15 | 南京大学 | Method for identifying plaque in blood vessel based on ultrasonic image |
CN114648514B (en) * | 2022-03-30 | 2022-11-29 | 中国人民解放军总医院第二医学中心 | Cerebral artery positioning and extracting method and device, electronic equipment and storage medium |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7397935B2 (en) * | 2004-05-10 | 2008-07-08 | Mediguide Ltd. | Method for segmentation of IVUS image sequences |
-
2014
- 2014-03-26 CN CN201410115141.8A patent/CN103886599B/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
Non-Patent Citations (2)
Title |
---|
Intravascular Ultrasound Image Hard Plaque Recognition and Media-adventitia Border Detection;XING Dong等;《Chinese Journal of Biomedical Engineering》;20120930(第03期);第110-116页 * |
血管内超声图像的预处理;汪友生等;《国外电子测量技术》;20120215;第31卷(第2期);第66-70页 * |
Also Published As
Publication number | Publication date |
---|---|
CN103886599A (en) | 2014-06-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103886599B (en) | A kind of blood vessel ROI dividing methods based on ivus image | |
CN108053417B (en) | lung segmentation device of 3D U-Net network based on mixed rough segmentation characteristics | |
WO2022083026A1 (en) | Ultrasound image denoising model establishing method and ultrasound image denoising method | |
CN107230206B (en) | Multi-mode data-based 3D pulmonary nodule segmentation method for hyper-voxel sequence lung image | |
CN104933709B (en) | Random walk CT lung tissue image automatic segmentation methods based on prior information | |
CN107909585B (en) | Intravascular intima segmentation method of intravascular ultrasonic image | |
WO2019001208A1 (en) | Segmentation algorithm for choroidal neovascularization in oct image | |
CN105741310B (en) | A kind of cardiac left-ventricle image segmenting system and method | |
CN112258514B (en) | Segmentation method of pulmonary blood vessels of CT (computed tomography) image | |
CN105303547A (en) | Multiphase CT image registration method based on grid matching Demons algorithm | |
Qin et al. | AirwayNet-SE: A simple-yet-effective approach to improve airway segmentation using context scale fusion | |
Shang et al. | Region competition based active contour for medical object extraction | |
CN107845065A (en) | Super-resolution image reconstruction method and device | |
CN111784701A (en) | Ultrasonic image segmentation method and system combining boundary feature enhancement and multi-scale information | |
CN112270667B (en) | TI-RADS-based integrated deep learning multi-tag identification method | |
CN106846330A (en) | Human liver's feature modeling and vascular pattern space normalizing method | |
CN109003280B (en) | Method for segmenting intima in blood vessel by two-channel intravascular ultrasonic image | |
CN113570627A (en) | Training method of deep learning segmentation network and medical image segmentation method | |
CN115147600A (en) | GBM multi-mode MR image segmentation method based on classifier weight converter | |
CN102663762B (en) | The dividing method of symmetrical organ in medical image | |
CN109472807B (en) | Blood vessel model extraction method based on deep neural network | |
CN111986216B (en) | RSG liver CT image interactive segmentation algorithm based on neural network improvement | |
CN117036253B (en) | Method for training a segmentation model for segmenting cerebral vessels and related products | |
CN106570880B (en) | In conjunction with the brain tissue MRI image dividing method of fuzzy clustering and markov random file | |
CN115731444A (en) | Medical image fusion method based on artificial intelligence and superpixel segmentation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171013 Termination date: 20200326 |