CN106910182A - In cardiac function MRI in end diastole image blood pool dividing method - Google Patents
In cardiac function MRI in end diastole image blood pool dividing method Download PDFInfo
- Publication number
- CN106910182A CN106910182A CN201510974234.0A CN201510974234A CN106910182A CN 106910182 A CN106910182 A CN 106910182A CN 201510974234 A CN201510974234 A CN 201510974234A CN 106910182 A CN106910182 A CN 106910182A
- Authority
- CN
- China
- Prior art keywords
- mri
- lamella
- blood pool
- region
- interest
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23211—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with adaptive number of clusters
-
- 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
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- 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
- G06T2207/30048—Heart; Cardiac
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention discloses a kind of dividing method of blood pool in cardiac function MRI end diastole image, it is comprised the following steps:Obtain some lamellas comprising myocardium of left ventricle and in the cardiac magnetic resonance images I of different phase phases aroused in interestNP；The MRI I corresponding to each phase phase aroused in interest in middle sheetMPOn, maximum intensity projection image of the initial area-of-interest in all phase phases aroused in interest is calculated, cluster segmentation maximum intensity projection image chooses the maximum region A of circularity；Calculate the MRI I corresponding to each phase phase aroused in interest of middle sheetMPImage averaging gray value in region a, the MRI I with maximum average gray valueMPIt is middle sheet in the corresponding MRI I of diastasisMP(ED)；MRI to the apex of the heart and heart bottom both direction lamella since middle sheet is split respectively, completes the MRI I of diastasisNP(ED)In blood pool segmentation.
Description
【Technical field】
The present invention relates to field of medical images, more particularly, to the segmentation of cardiac function MRI.
【Background technology】
In recent years, the morbidity and mortality of angiocardiopathy just increase year by year.Have nearly 600 in the U.S. every year
Ten thousand people are checked out with angiocardiopathy, and China then has nearly 3,000,000 people to die from angiocardiopathy.Heart
Function is the important indicator of diagnosis of cardiovascular diseases, can be by left in the four-dimensional heart function MRI of analysis
The related global and local characteristic parameter realization of ventricle, such as：Ventricular volume, LVEF and myocardium wall thickness etc..
The accuracy of heart function characteristic parameter extraction depends on being accurately positioned for the intracardiac adventitia of left ventricle, and computer aided manufacturing
Help diagnosis scheme be favorably improved in four-dimensional heart function MRI position the intracardiac adventitia of left ventricle can
By property and efficiency.
Determine that end diastole is typically manually and needs what user aided in four-dimensional heart function MRI
Semi-automatic method, this kind of method is more complicated, efficiency is low and exists between larger observer and observes
The difference of person itself.Relaxed in the computer-aided diagnosis scheme cardiac based on cardiac function MRI
The segmentation of blood pool is firstly the need of the major issue for solving in the determination in latter stage and end diastole image.
【The content of the invention】
The technical problems to be solved by the invention are to provide diastasis in a kind of cardiac function MRI
The full-automatic partition method of blood pool in image.
The present invention is to solve above-mentioned technical problem and the technical scheme that uses is：A kind of cardiac function magnetic resonance
The dividing method of blood pool, comprises the following steps in image end diastole image:
Obtain some lamellas comprising myocardium of left ventricle and in the cardiac magnetic resonance images I of different phase phases aroused in interestNP,
Wherein N represents the sequence number of lamella, and P represents phase phase sequence number aroused in interest, and N, P are the integer more than or equal to 1；
The MRI I corresponding to each phase phase aroused in interest in middle sheetMPOn, set initial region of interest
Domain, calculates maximum intensity projection image of the initial area-of-interest in all phase phases aroused in interest, and cluster segmentation is most
Big intensity projection images, choose the maximum region A of circularity；
Calculate the MRI I corresponding to each phase phase aroused in interest of middle sheetMPImage in region a is put down
Equal gray value, the MRI I with maximum average gray valueMPThe place phase aroused in interest is mutually defined as diastole
Latter stage, and determine middle sheet in the corresponding MRI I of diastasisMP(ED)；
MRI to the apex of the heart and heart bottom both direction lamella since middle sheet is split respectively,
Complete the MRI I of diastasisNP(ED)In blood pool segmentation.
Preferably, including left ventricle area-of-interest is obtained, the center of the left ventricle area-of-interest is area
The barycenter of domain A, the length of side of the left ventricle area-of-interest is the long axis length addend pixel of region A.
Preferably, to the MRI I of the diastasis of middle sheetMP(ED)In blood pool segmentation include with
Lower step：
To the MRI I of the diastasis of middle sheetMP(ED)Image in left ventricle area-of-interest enters
Row gamma correction, the image I after being correctedMP(ED)’；
By fuzzy C-means clustering method to the image I after correctionMP(ED)' classified, classification number is 2,
Most bright class is chosen as candidate region SM；
In candidate region SMThe middle diastasis for choosing the region maximum with region A Duplication as middle sheet
MRI IMP(ED)In blood pool region segmentation result.
Preferably, to the blood pool and M+1 of first to the M-1 MRI of the diastasis of lamella
The blood pool segmentation of the MRI of the individual diastasis to n-th lamella is obtained in the following manner：
The magnetic resonance of current lamella is guided with the segmentation result in the blood pool region of the MRI of previous lamella
The segmentation in the blood pool region of image.
Preferably, it is current lamella to set K-1, K+1 lamella, and k-th lamella is previous lamella；
K-1, K+1 MRI I of the diastasis of lamella is obtained according to following steps(K+1)P(ED)、I(K-1) P(ED)Blood pool region segmentation：
With k-th lamella, the MRI I of diastasisKP(ED)In split the barycenter and length in blood pool region
Axle addend pixel as K+1, the K-1 lamella, diastasis MRI I(K+1)P(ED)、
I(K-1)P(ED)Left ventricle area-of-interest center and the length of side, and to the image in left ventricle area-of-interest
Gamma correction is carried out, the image I after being corrected(K+1)P(ED)’、I(K-1)P(ED)’；
By fuzzy C-means clustering method to the image I after correction(K+1)P(ED)’、I(K-1)P(ED)' divided
Class, chooses most bright class as candidate region S(K-1)、S(K+1)；
In K-1, K+1 candidate region S of lamella(K-1)、S(K+1)Middle selection and k-th lamella, diastole
The MRI I in latter stageKP(ED)The maximum region of the segmentation result Duplication in middle blood pool region as K-1,
The segmentation result in blood pool region in K+1 lamella, the MRI of diastasis.
Preferably, if the major axis in the blood pool region split in the MRI of current lamella is more than previous lamella
MRI in split 1.2 times of major axis of blood pool region or current lamella MRI
The length-width ratio in the blood pool region of middle segmentation is more than the blood pool region split in the MRI of previous lamella
1.3 times of length-width ratio, then be judged as LVOT phenomenons occur, wherein current slice layer is to be located at centre
Lamella below slice position.
Preferably, if there is LVOT phenomenons, then gamma correction parameter, the magnetic resonance to current lamella are adjusted
Image in the left ventricle area-of-interest of image carries out gamma correction, and re-starts cluster segmentation.
Preferably, as using adjustment gamma correction parameter after and carry out the current lamella after cluster segmentation magnetic be total to
The blood pool region shaken in image is still judged as LVOT phenomenons occur, then using the removal of ray scanning method
The region of over-segmentation.
Preferably, the ray scanning method is comprised the following steps：
Point makees radius scanning centered on the blood pool region barycenter of the MRI of current lamella, extracts
The nearest zone boundary point of distance center point on each angle ray；
All boundary points of said extracted to the radial distance average mean and standard deviation std of central point are calculated,
Remove zone boundary point of the radial distance more than mean+std；
Closed curve is obtained after remaining zone boundary is clicked through into row interpolation, the region that closed curve is included is
The blood pool region of the MRI of current lamella.
Preferably, the center of the initial area-of-interest is the center of original image, and radius is 100 pixels
Point.
Present invention contrast prior art has following beneficial effect：Using the method for the present invention, it is capable of achieving entirely certainly
The dynamic phase phase and left ventricular location for determining late diastole, end diastole image is realized using clustering method
The segmentation of middle blood pool.
【Brief description of the drawings】
Fig. 1 is blood in end diastole image in a kind of cardiac function MRI of the embodiment of the present invention
The flow chart of the dividing method in pond；
Fig. 2 a are middle sheet MRI maximum intensity projection's image in the region of interest；
Fig. 2 b are the image in initial area-of-interest of middle sheet MRI；
Fig. 2 c are the result for carrying out cluster segmentation to Fig. 2 b by fuzzy C-means clustering method；
Fig. 2 d are the result of blood pool region segmentation in middle sheet MRI；
The process schematic that Fig. 3 a-3c are processed overdivided region by ray scanning method.
【Specific embodiment】
Fig. 1-3 are referred to, a kind of cardiac function MRI diastasis figure in the embodiment of the present invention
The dividing method of blood pool, comprises the following steps as in:
Obtain some lamellas comprising myocardium of left ventricle and in the cardiac magnetic resonance images I of different phase phases aroused in interestNP,
Wherein N represents the sequence number of lamella, and P represents phase phase (phase) sequence number aroused in interest, and N, P are more than or wait
In 1 integer；
The MRI I corresponding to each phase phase aroused in interest in middle sheetMPOn, set initial region of interest
Domain, calculates maximum intensity projection of the initial area-of-interest on all phase phases (P phase phase aroused in interest) aroused in interest
Image (MIP), cluster segmentation maximum intensity projection image (MIP) chooses the maximum region A of circularity；It is excellent
The cluster segmentation method of choosing is fuzzy C-means clustering method；
Calculate the MRI I corresponding to each phase phase aroused in interest of middle sheetMPImage in region a is put down
Equal gray value, the MRI I with maximum average gray valueMPThe place phase aroused in interest is mutually defined as diastole
Latter stage P (ED), and determine middle sheet in the corresponding MRI I of diastasisMP(ED)；
MRI to the apex of the heart and heart bottom both direction lamella since middle sheet is split respectively,
Complete the MRI I of (N number of lamella) diastasisNP(ED)In blood pool segmentation.
Phase aroused in interest refers to mutually that will be divided since the complete cardiac cycle T the systole phase of left ventricle to diastasis
It is multiple time periods (phase) that these time periods can be one to be the continuous time series of class, or
Interlude sequence.
Several lamellas (slice) are obtained along the long axis direction of heart in region of the ventricle comprising muscle portion,
And each lamella, the cardiac magnetic resonance images I in each phase phase aroused in interest are obtained by magnetic resonance imagingNP,
Wherein N represents the sequence number of corresponding lamella, and P represents phase phase (phase) sequence number aroused in interest, and N, P are and are more than
Or the integer equal to 1；Specifically, the value of N can be 3,5,6 etc., the value of P can for 3,5,
8th, 10 etc.；In addition, it is middle sheet to define m-th lamella, wherein, M is whole more than or equal to 2
Number.
The each lamella and cardiac magnetic resonance images in each phase phase are to obtain in the following manner：
By the MR data line in collection multiple cardiac cycles, corresponding same cardiac phase phases, will be adopted
The MR data line for collecting is filled into a K-Space (K spaces), and phase is obtained by Fourier transformation
Lamella is answered in the MRI of corresponding phase phase aroused in interest.By setting less time window, such as 20-50ms,
And in multiple cardiac cycle gathered datas, influence of the cardiac motion artefacts to image can be reduced, improve image
Quality.
Initial region of interest ROI0Center be original image (i.e. MRI IMP) center, radius is
Several pixels, the size of radius is 60-120 pixel, and preferably radius size is 100 pixels
Point.
Then, maximum intensity projection image (MIP) of the initial area-of-interest in all phase phases aroused in interest is calculated,
Maximum intensity projection's image (MIP) is split by fuzzy C-means clustering method, the maximum region of circularity is chosen
A；Wherein, the number of cluster (classification) can be 2,3 or 4 etc.；The definition of circularity is (girth * weeks
It is long)/(4*PI* areas).
Further, left ventricle area-of-interest is obtained, the barycenter of the left ventricle area-of-interest is region
The barycenter of A, the length of side of the left ventricle area-of-interest is the long axis length addend pixel of region A,
Concretely 10-30 pixel, preferably 20 pixels.
Further, to the MRI I of the diastasis of middle sheetMP(ED)In the segmentation of blood pool include
Following steps：
To the MRI I of the diastasis of middle sheetMP(ED)Image in left ventricle area-of-interest enters
Row gamma correction (Gamma Correction), the image I after being correctedMP(ED)’；By gamma correction,
The contrast of blood pool and cardiac muscle can be strengthened, facilitate follow-up identification and segmentation；
By fuzzy C-means clustering method to the image I after correctionMP(ED)' classified, classification number is 2,
Most bright class is chosen as candidate region SM；
In candidate region SMThe middle diastasis for choosing the region maximum with region A Duplication as middle sheet
MRI IMP(ED)In blood pool region (Blood Pool, BP) segmentation result.
Further, to first to the M-1 blood pool segmentation of the MRI of the diastasis of lamella,
Or the blood pool segmentation of the MRI of the M+1 diastasis to n-th lamella is obtained in the following manner
：The magnetic resonance of current lamella is guided with the segmentation result in the blood pool region of the MRI of previous lamella
The segmentation in the blood pool region of image.
If specifically, M-1 lamella or the M+1 lamella are used as current lamella, m-th lamella is then
Its previous lamella, the M-1 lamella is then the M-2 previous lamella of lamella, the M+1 lamella
It is then the M+2 previous lamella of lamella, the like.
Specifically, needing first in m-th lamella, the MRI I of diastasisMP(ED)Middle blood pool region is complete
Into after segmentation, with m-th lamella, the MRI I of diastasisMP(ED)The segmentation knot in middle blood pool region
On the basis of fruit, to M-1, the M+1 lamella, diastasis MRI I(M-1)P(ED)、I(M+1) P(ED)Middle blood pool region is split, then with the M-1 lamella, the MRI I of diastasis(M-1) P(ED)Blood pool region segmentation result on the basis of, to the M-2 lamella, the MRI I of diastasis(M-2)P(ED)Blood pool region split；Or, then with the M+1 lamella, the magnetic resonance of diastasis
Image I(M+1)P(ED)Blood pool region segmentation result on the basis of, to the M+2 lamella, the magnetic of diastasis
Resonance image I(M+2)P(ED)Blood pool region split, the like.
It is current lamella to set K-1, K+1 lamella, and k-th lamella is previous lamella, and k-th
The MRI I of lamella, diastasisKP(ED)Blood pool completed segmentation；Obtained according to following steps
K-1, K+1 MRI I of the diastasis of lamella(K+1)P(ED)、I(K-1)P(ED)Blood pool region
Segmentation：
With previous (K) lamella, the MRI I of diastasisKP(ED)In split the matter in blood pool region
The heart and major axis addend (such as 20-60) pixel are used as K+1, the K-1 lamella, end-diastolic
The MRI I of phase(K+1)P(ED)、I(K-1)P(ED)Left ventricle area-of-interest center and the length of side, it is and right
Image in left ventricle area-of-interest carries out gamma correction, the image I after being corrected(K+1)P(ED)’、I(K-1) P(ED)’；
By fuzzy C-means clustering method to the image I after correction(K+1)P(ED)’、I(K-1)P(ED)' divided
Class, chooses most bright class as candidate region S(K-1)、S(K+1)；
In K-1, K+1 candidate region S of lamella(K-1)、S(K+1)Middle selection and k-th lamella, diastole
The MRI I in latter stageKP(ED)The maximum region of the segmentation result Duplication in middle blood pool region as K-1,
The segmentation result in blood pool region in K+1 lamella, the MRI of diastasis.
Further, if the major axis in the blood pool region split in the MRI of current lamella is more than preceding a piece of
1.2 times of the major axis in the blood pool region split in the MRI of layer or the magnetic resonance figure of current lamella
The length-width ratio in the blood pool region split as in is more than the blood pool region split in the MRI of previous lamella
1.3 times of length-width ratio, then be judged as LVOT (left ventricular outflow tract) occur
Phenomenon, wherein current slice layer is to be located at the lamella below middle slice position (near heart bottom).
Further, if there is LVOT phenomenons, then gamma correction parameter (increase correction parameter) is adjusted,
Gamma correction is carried out to the image in the left ventricle area-of-interest of the MRI of current lamella, and again
Carry out cluster segmentation.
Further, as using after adjustment gamma correction parameter and carrying out the magnetic of the current lamella after cluster segmentation
Blood pool region in resonance image is still judged as LVOT phenomenons occur, then using ray scanning (ray
Scanning) method removes the region of over-segmentation, you can the blood pool region after being split.
Specifically, the ray scanning method is comprised the following steps：
The point a centered on the blood pool region barycenter of the MRI of current lamella (ginseng Fig. 3 a) makees radially to penetrate
Line is scanned, and extracts the nearest zone boundary point of distance center point on each the angle ray (point in ginseng Fig. 3 a
b1、b2、b3)；
All boundary points of said extracted to the radial distance average mean and standard deviation std of central point are calculated,
Remove radial distance more than apart from zone boundary point (b3) of average (mean) plus standard deviation (std), protect
The point for staying is as shown in Figure 3 b；
Closed curve d (ginseng Fig. 3 c) is obtained after remaining zone boundary is clicked through into row interpolation；Closed curve is included
Region be current lamella MRI blood pool region.
It should be noted that through the above description of the embodiments, those skilled in the art can understand
Recognize that the part or all of of the application can be realized by software and the required general hardware platform of combination in ground.
Based on such understanding, the part that the technical scheme of the application substantially contributes to prior art in other words
Can be embodied in the form of software product, the computer software product may include to be stored thereon with machine can
One or more machine readable medias of execute instruction, these instructions are by such as computer, computer network
Or one or more machines such as other electronic equipments may be such that one or more machines according to this hair when performing
Bright embodiment performs operation.Machine readable media may include, but be not limited to, floppy disk, CD, CD-ROM
(compact-disc-read-only storage), magneto-optic disk, ROM (read-only storage), RAM (random access memory),
EPROM (Erasable Programmable Read Only Memory EPROM), EEPROM (Electrically Erasable Read Only Memory),
Magnetic or optical card, flash memory are suitable to store the other kinds of medium/machine readable of machine-executable instruction
Medium.
The application can be used in numerous general or special purpose computing system environments or configuration.For example：Individual calculus
Machine, server computer, handheld device or portable set, laptop device, multicomputer system, base
System, set top box, programmable consumer-elcetronics devices, network PC, minicom in microprocessor,
Mainframe computer, including the DCE of any of the above system or equipment etc..
The application can be described in the general context of computer executable instructions, example
Such as program module.Usually, program module includes performing particular task or realizes particular abstract data type
Routine, program, object, component, data structure etc..This can also in a distributed computing environment be put into practice
Application, in these DCEs, by the remote processing devices connected by communication network come
Execution task.In a distributed computing environment, program module may be located at including local including storage device
In remote computer storage medium.
Using the method for the present invention, full-automatic determination late diastole phase and left ventricular location are capable of achieving,
The segmentation of blood pool in end diastole image is realized using clustering method.It is total to based on cardiac function magnetic for follow-up
The computer-aided diagnosis of image of shaking provides strong support and guarantee, and improves the efficiency and standard of follow-up diagnosis
True property.
Although the present invention is disclosed as above with preferred embodiment, so it is not limited to the present invention, Ren Heben
Art personnel, without departing from the spirit and scope of the present invention, when a little modification and perfect can be made,
Therefore protection scope of the present invention is when by being defined that claims are defined.
Claims (10)
1. in a kind of cardiac function MRI end diastole image blood pool dividing method, its feature exists
In comprising the following steps:
Obtain some lamellas comprising myocardium of left ventricle and in the cardiac magnetic resonance images I of different phase phases aroused in interestNP,
Wherein N represents the sequence number of lamella, and P represents phase phase sequence number aroused in interest, and N, P are the integer more than or equal to 1；
The MRI I corresponding to each phase phase aroused in interest in middle sheetMPOn, set initial region of interest
Domain, calculates maximum intensity projection image of the initial area-of-interest in all phase phases aroused in interest, and cluster segmentation is most
Big intensity projection images, choose the maximum region A of circularity；
Calculate the MRI I corresponding to each phase phase aroused in interest of middle sheetMPImage in region a is put down
Equal gray value, the MRI I with maximum average gray valueMPThe place phase aroused in interest is mutually defined as diastole
Latter stage, and determine middle sheet in the corresponding MRI I of diastasisMP(ED)；
MRI to the apex of the heart and heart bottom both direction lamella since middle sheet is split respectively,
Complete the MRI I of diastasisNP(ED)In blood pool segmentation.
2. blood pool in end diastole image in cardiac function MRI according to claim 1
Dividing method, it is characterised in that including obtaining left ventricle area-of-interest, the left ventricle area-of-interest
Center be the barycenter of region A, the length of side of the left ventricle area-of-interest adds for the long axis length of region A
Several pixels.
3. blood pool in end diastole image in cardiac function MRI according to claim 2
Dividing method, it is characterised in that to the MRI I of the diastasis of middle sheetMP(ED)In blood pool
Segmentation comprise the following steps：
To the MRI I of the diastasis of middle sheetMP(ED)Image in left ventricle area-of-interest enters
Row gamma correction, the image I after being correctedMP(ED)’；
By fuzzy C-means clustering method to the image I after correctionMP(ED)' classified, classification number is 2,
Most bright class is chosen as candidate region SM；
In candidate region SMThe middle diastasis for choosing the region maximum with region A Duplication as middle sheet
MRI IMP(ED)In blood pool region segmentation result.
4. blood pool in end diastole image in cardiac function MRI according to claim 3
Dividing method, it is characterised in that to first to the M-1 MRI of the diastasis of lamella
The blood pool of the MRI of blood pool and the M+1 diastasis to n-th lamella is split by with lower section
Formula is obtained：
The magnetic resonance of current lamella is guided with the segmentation result in the blood pool region of the MRI of previous lamella
The segmentation in the blood pool region of image.
5. blood pool in end diastole image in cardiac function MRI according to claim 4
Dividing method, it is characterised in that it is current lamella to set K-1, K+1 lamella, and k-th lamella is
Previous lamella；K-1, K+1 MRI of the diastasis of lamella is obtained according to following steps
I(K+1)P(ED)、I(K-1)P(ED)Blood pool region segmentation：
With k-th lamella, the MRI I of diastasisKP(ED)In split the barycenter and length in blood pool region
Axle addend pixel as K+1, the K-1 lamella, diastasis MRI I(K+1)P(ED)、
I(K-1)P(ED)Left ventricle area-of-interest center and the length of side, and to the image in left ventricle area-of-interest
Gamma correction is carried out, the image I after being corrected(K+1)P(ED)’、I(K-1)P(ED)’；
By fuzzy C-means clustering method to the image I after correction(K+1)P(ED)’、I(K-1)P(ED)' divided
Class, chooses most bright class as candidate region S(K-1)、S(K+1)；
In K-1, K+1 candidate region S of lamella(K-1)、S(K+1)Middle selection and k-th lamella, diastole
The MRI I in latter stageKP(ED)The maximum region of the segmentation result Duplication in middle blood pool region as K-1,
The segmentation result in blood pool region in K+1 lamella, the MRI of diastasis.
6. blood pool in end diastole image in cardiac function MRI according to claim 5
Dividing method, it is characterised in that if the major axis in the blood pool region split in the MRI of current lamella is big
1.2 times of major axis of the blood pool region split in the MRI of previous lamella or current lamella
The length-width ratio in the blood pool region split in MRI is more than segmentation in the MRI of previous lamella
1.3 times of the length-width ratio in blood pool region, then be judged as LVOT phenomenons occur, wherein current slice layer
It is to be located at the lamella below middle slice position.
7. blood pool in end diastole image in cardiac function MRI according to claim 6
Dividing method, it is characterised in that if there is LVOT phenomenons, then gamma correction parameter is adjusted, to current
Image in the left ventricle area-of-interest of the MRI of lamella carries out gamma correction, and re-starts poly-
Class is split.
8. blood pool in end diastole image in cardiac function MRI according to claim 7
Dividing method, it is characterised in that as using after adjustment gamma correction parameter and carrying out current after cluster segmentation
Blood pool region in the MRI of lamella is still judged as LVOT phenomenons occur, then swept using ray
Retouch the region that method removes over-segmentation.
9. blood pool in end diastole image in cardiac function MRI according to claim 8
Dividing method, it is characterised in that the ray scanning method is comprised the following steps：
Point makees radius scanning centered on the blood pool region barycenter of the MRI of current lamella, extracts
The nearest zone boundary point of distance center point on each angle ray；
All boundary points of said extracted to the radial distance average mean and standard deviation std of central point are calculated,
Remove zone boundary point of the radial distance more than mean+std；
Closed curve is obtained after remaining zone boundary is clicked through into row interpolation, the region that closed curve is included is
The blood pool region of the MRI of current lamella.
10. blood pool in end diastole image in cardiac function MRI according to claim 1
Dividing method, it is characterised in that the center of the initial area-of-interest for original image center, radius
It is 100 pixels.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510974234.0A CN106910182B (en) | 2015-12-22 | 2015-12-22 | Segmentation method of blood pool in end diastole image in cardiac function magnetic resonance image |
| US15/387,758 US10290109B2 (en) | 2015-12-22 | 2016-12-22 | Method and system for cardiac image segmentation |
| US16/372,526 US10776924B2 (en) | 2015-12-22 | 2019-04-02 | Method and system for cardiac image segmentation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510974234.0A CN106910182B (en) | 2015-12-22 | 2015-12-22 | Segmentation method of blood pool in end diastole image in cardiac function magnetic resonance image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106910182A true CN106910182A (en) | 2017-06-30 |
| CN106910182B CN106910182B (en) | 2020-11-27 |
Family
ID=59200230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510974234.0A Active CN106910182B (en) | 2015-12-22 | 2015-12-22 | Segmentation method of blood pool in end diastole image in cardiac function magnetic resonance image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106910182B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109001660A (en) * | 2018-06-12 | 2018-12-14 | 上海联影医疗科技有限公司 | Film imaging method and magnetic resonance imaging system |
| CN113362345A (en) * | 2021-06-30 | 2021-09-07 | 武汉中科医疗科技工业技术研究院有限公司 | Image segmentation method and device, computer equipment and storage medium |
| CN115965621A (en) * | 2023-02-15 | 2023-04-14 | 中国医学科学院阜外医院 | Main heart adverse event prediction device based on magnetic resonance imaging |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100215238A1 (en) * | 2009-02-23 | 2010-08-26 | Yingli Lu | Method for Automatic Segmentation of Images |
| CN103340628A (en) * | 2013-06-28 | 2013-10-09 | 中国科学院深圳先进技术研究院 | Method and system for processing heart real-time film imaged picture |
| CN104519236A (en) * | 2013-09-26 | 2015-04-15 | 京瓷办公信息系统株式会社 | Image forming apparatus and image forming method |
| CN104978730A (en) * | 2014-04-10 | 2015-10-14 | 上海联影医疗科技有限公司 | Division method and device of left ventricular myocardium |
-
2015
- 2015-12-22 CN CN201510974234.0A patent/CN106910182B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100215238A1 (en) * | 2009-02-23 | 2010-08-26 | Yingli Lu | Method for Automatic Segmentation of Images |
| CN103340628A (en) * | 2013-06-28 | 2013-10-09 | 中国科学院深圳先进技术研究院 | Method and system for processing heart real-time film imaged picture |
| CN104519236A (en) * | 2013-09-26 | 2015-04-15 | 京瓷办公信息系统株式会社 | Image forming apparatus and image forming method |
| CN104978730A (en) * | 2014-04-10 | 2015-10-14 | 上海联影医疗科技有限公司 | Division method and device of left ventricular myocardium |
Non-Patent Citations (3)
| Title |
|---|
| CAROLINE PETITJEAN.ET AL: "《A review of segmentation methods in short axis cardiac MR images》", 《MEDICAL IMAGE ANALYSIS》 * |
| HUAIFEI HU.ET AL: "《Hybrid segmentation of left ventricle in cardiac MRI using gaussian-mixture model and region restricted dynamic programming》", 《MAGNETIC RESONANCE IMAGING》 * |
| YING-LI LU.ET AL: "《Automatic functional analysis of left ventricle in cardiac cine MRI》", 《QUANTITATIVE IMAGING IN MEDICINE AND SURGERY》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109001660A (en) * | 2018-06-12 | 2018-12-14 | 上海联影医疗科技有限公司 | Film imaging method and magnetic resonance imaging system |
| CN109001660B (en) * | 2018-06-12 | 2020-07-28 | 上海联影医疗科技有限公司 | Film imaging method and magnetic resonance imaging system |
| CN113362345A (en) * | 2021-06-30 | 2021-09-07 | 武汉中科医疗科技工业技术研究院有限公司 | Image segmentation method and device, computer equipment and storage medium |
| CN115965621A (en) * | 2023-02-15 | 2023-04-14 | 中国医学科学院阜外医院 | Main heart adverse event prediction device based on magnetic resonance imaging |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106910182B (en) | 2020-11-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10776924B2 (en) | Method and system for cardiac image segmentation | |
| JP7383010B2 (en) | Medical image recognition method and system, model training method, computer device, and program | |
| JP6884853B2 (en) | Image segmentation using neural network method | |
| JP6932182B2 (en) | Systems and methods for image segmentation with convolutional neural networks | |
| EP2916738B1 (en) | Lung, lobe, and fissure imaging systems and methods | |
| JP4918048B2 (en) | Image processing apparatus and method | |
| JP5438029B2 (en) | Automatic 3D segmentation of short axis delay-enhanced cardiac MRI | |
| EP3633612A1 (en) | Method and system for segmenting image | |
| EP2549437B1 (en) | Moving object contour extraction apparatus, left ventricle image segmentation apparatus, moving object contour extraction method and left ventricle image segmentation | |
| US9881360B2 (en) | Multi-scale active contour segmentation | |
| CN104424629B (en) | A kind of x-ray chest radiograph lung segmentation method and apparatus | |
| JP6273291B2 (en) | Image processing apparatus and method | |
| US8948484B2 (en) | Method and system for automatic view planning for cardiac magnetic resonance imaging acquisition | |
| CN106910194A (en) | The dividing method of the intracardiac adventitia in cardiac function MRI | |
| JP4991697B2 (en) | Method, system and computer program for partitioning structures in a data set | |
| Mahmoudi et al. | GPU-based segmentation of cervical vertebra in X-ray images | |
| JP5709216B2 (en) | Image processing program, method and apparatus | |
| Larrey-Ruiz et al. | Automatic image-based segmentation of the heart from CT scans | |
| Lei et al. | Echocardiographic image multi‐structure segmentation using Cardiac‐SegNet | |
| CN106910182A (en) | In cardiac function MRI in end diastole image blood pool dividing method | |
| Medina et al. | Left ventricle myocardium segmentation in multi-slice computerized tomography | |
| Alizadeh Sani et al. | Automatic computation of left ventricular volume changes over a cardiac cycle from echocardiography images by nonlinear dimensionality reduction | |
| O'Dell et al. | Traversing and labeling interconnected vascular tree structures from 3D medical images | |
| Alom et al. | Automatic slice growing method based 3D reconstruction of liver with its vessels | |
| US20230030618A1 (en) | Making measurements in 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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 201807 Shanghai city Jiading District Industrial Zone Jiading Road No. 2258 Applicant after: Shanghai Lianying Medical Technology Co., Ltd Address before: 201807 Shanghai city Jiading District Industrial Zone Jiading Road No. 2258 Applicant before: SHANGHAI UNITED IMAGING HEALTHCARE Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |