CN105741310B - A kind of cardiac left-ventricle image segmenting system and method - Google Patents
A kind of cardiac left-ventricle image segmenting system and method Download PDFInfo
- Publication number
- CN105741310B CN105741310B CN201610161395.2A CN201610161395A CN105741310B CN 105741310 B CN105741310 B CN 105741310B CN 201610161395 A CN201610161395 A CN 201610161395A CN 105741310 B CN105741310 B CN 105741310B
- Authority
- CN
- China
- Prior art keywords
- left ventricle
- profile
- image
- sustainer
- minimum
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- 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
- G06T2207/20036—Morphological image processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20116—Active contour; Active surface; Snakes
-
- 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
- Image Analysis (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The present invention provides a kind of cardiac left-ventricle image segmenting system and method.The system includes：Image conversion unit；Left ventricle profile coarse extraction unit；Left ventricle profile essence extraction unit.This method includes：Segmentation nuclear magnetic resonance image obtains binary image, and binary image is converted into gray level image by Euclidean Distance Transform；Each linking area in binary image is spaced apart, coarse extraction goes out left ventricle profile；Judge whether left ventricle profile is connected with sustainer profile, if the left ventricle profile that coarse extraction goes out is connected with sustainer profile, removes aorta segmentation and open, repair left ventricle profile, obtain cardiac left-ventricle image segmentation result；Otherwise the left ventricle profile that coarse extraction goes out is used as left ventricle profile.The present invention can exclude the influence that left ventricle is connected with sustainer in terms of image, accurately split left ventricle bottom layer image, the influence of sustainer edge leakage caused by left ventricle is overcome, so as to obtain accurate left ventricle segmentation result.
Description
Technical field
The present invention relates to image processing field, and in particular to a kind of cardiac left-ventricle image segmenting system and method.
Background technology
In numerous image partition methods, there are many methods for being adapted to segmentation cardiac magnetic resonance short axis images, as region is given birth to
Long method, threshold method and Level Set Method etc..Such as in paper " outside the MR image left ventricles of shape Statistics Mumford-Shah models
The Mumford-Shah model dividing methods employed in contours segmentation " with reference to shape Statistics split heart left ventricle,
The situation of weak boundary and boundary fault is considered.But it have ignored heart that bottom causes portion due to sustainer
The situation for dividing left ventricle border to be not present.And paper " with reference to texture and the Tagged MR image left ventricles partitioning algorithm of shape "
Texture classification information and statistical shape prior knowledge are introduced into Mumford-Shah models, it is proposed that a kind of improved segmentation band
The method of the left ventricle inside and outside contour of mark line nuclear magnetic resonance image, situation about not being connected to heart bottom with sustainer equally add
To consider, it is difficult to differentiate between on texture, does not also introduce solution method specifically in shape Statistics." improved in paper based on Snake
In the cardiac MR images left ventricle dividing method of model ", the main segmentation for focusing on outer membrane in the left ventricle of heart middle level, not to heart
The situation of bottom edges leakage takes in.
Heart bottom causes edge leakage due to being connected with sustainer, so segmentation is got up, difficulty is bigger, greatly
Partial segmentation method is all difficult to handle the problem of such.Most of scholar mainly focuses on to middle level when splitting heart left ventricle
Segmentation, ignores the processing to left ventricle bottom, but heart left ventricle's bottom is connected to sustainer, to heart left ventricle's bottom
Processing is to obtain the various information of heart left ventricle and to necessary to heart left ventricle's progress three-dimensional reconstruction.
The content of the invention
In view of the deficienciess of the prior art, the present invention provides a kind of cardiac left-ventricle image segmenting system and method.
The technical scheme is that：
A kind of cardiac left-ventricle image segmenting system, including：
Image conversion unit：The selected seed point on the left ventricle of nuclear magnetic resonance image, by the maximum of nuclear magnetic resonance image
Nuclear magnetic resonance image is divided into foreground and background, incited somebody to action by gray value and the average value of minimum gradation value as initial segmentation threshold value
The average gray value of prospect and the average gray value of background take it is average as new segmentation threshold, if new segmentation threshold with it is previous
The difference of segmentation threshold obtained by iteration is less than the allowed band of setting, then current segmentation threshold is final segmentation threshold, otherwise
Continue to iterate to calculate, finally obtain binary image, then binary image is converted into by gray level image by Euclidean Distance Transform；
Left ventricle profile coarse extraction unit：Sort by pixel of the incremental order of gray scale to gray level image, use one
FIFO queues recursively distribute to each gray scale very small region in the way of breadth-first and form the minimum area of new gray scale respectively
Domain, and the pixel of first gray scale very small region is collectively labeled as 0 in distribution, the mark of gray scale very small region below according to
It is secondary to add 1, each linking area in gray level image is distinguished by mark, left ventricle is gone out by the seed point coarse extraction of selection
Profile；
Left ventricle profile essence extraction unit：If the left ventricle profile that coarse extraction goes out is connected with sustainer profile, master is removed
Artery, repairs left ventricle profile, obtains cardiac left-ventricle image segmentation result；If left ventricle profile and sustainer that coarse extraction goes out
Profile is not attached to, then the left ventricle profile that coarse extraction goes out is used as left ventricle profile, obtains cardiac left-ventricle image segmentation result.
The left ventricle profile essence extraction unit, including：
Analyze judging unit：The left ventricle profile barycenter that calculating coarse extraction goes out finds out distance most to the distance of each marginal point
Big value Ma and apart from minimum M i, and according to apart from maximum and apart from minimum value analysis and judge left ventricle whether with sustainer
It is connected：If Ma<30 pixel unit lengths and (Ma-Mi)/Ma>0.55, or Ma>30 pixel unit lengths and
(Ma-Mi)/Ma>0.38, then left ventricle be connected with sustainer, it is necessary to carry out Morphological scale-space, the left ventricle that otherwise coarse extraction goes out
Profile is used as left ventricle profile.
The left ventricle profile essence extraction unit, further includes：
Morphological scale-space unit：Barycenter in the left ventricle profile that coarse extraction goes out is found out to all minimums of marginal point, if
Apart from the position of maximum Ma between two adjacent minimum points, then making straight line as cut-point using the two minimum points will
Left ventricle profile is separated with sustainer profile, redefines the barycenter of left ventricle profile, in two cut-points and the barycenter phase
Even the center line direction of angulation takes the distance average of two cut-points to barycenter to make the 3rd cut-point, is split with these three
Point makees curve completion left ventricle contour edge, obtains left ventricle profile.
The minimum point Pi meets (Pi-Mi)/(Ma-Mi)<0.3.
The present invention also provides a kind of cardiac left-ventricle image dividing method, including：
Image is changed：The selected seed point on the left ventricle of nuclear magnetic resonance image, by the maximum gray scale of nuclear magnetic resonance image
Nuclear magnetic resonance image is divided into foreground and background, by prospect by the average value of value and minimum gradation value as initial segmentation threshold value
Average gray value and background average gray value take it is average as new segmentation threshold, if new segmentation threshold and previous iteration
The difference of the segmentation threshold of gained is less than the allowed band of setting, then current segmentation threshold is final segmentation threshold, is otherwise continued
Iterative calculation, finally obtains binary image, then binary image is converted into gray level image by Euclidean Distance Transform；
Coarse extraction left ventricle profile：Sort by pixel of the incremental order of gray scale to gray level image, use a FIFO team
Row recursively distribute to each gray scale very small region in the way of breadth-first and form new gray scale very small region respectively, and
The pixel of first gray scale very small region is collectively labeled as 0 during distribution, the mark of gray scale very small region below adds 1 successively,
Each linking area in gray level image is distinguished by mark, left ventricle profile is gone out by the seed point coarse extraction of selection；
Essence extraction left ventricle profile：If the left ventricle profile that coarse extraction goes out is connected with sustainer profile, sustainer is removed,
Left ventricle profile is repaired, obtains cardiac left-ventricle image segmentation result；If the left ventricle profile that coarse extraction goes out and sustainer profile
It is not attached to, then the left ventricle profile that coarse extraction goes out is used as left ventricle profile, obtains cardiac left-ventricle image segmentation result.
The specific steps of the essence extraction left ventricle profile include：
Distance of the left ventricle profile barycenter that calculating coarse extraction goes out to each marginal point；
Find out apart from maximum Ma and apart from minimum M i；
According to analyzing apart from maximum and apart from minimum value and judge whether left ventricle is connected with sustainer：If Ma<30
A pixel unit length and (Ma-Mi)/Ma>0.55, or Ma>30 pixel unit lengths and (Ma-Mi)/Ma>0.38,
Then left ventricle obtains left ventricle profile after being connected with sustainer, it is necessary to carry out Morphological scale-space, the left ventricle that otherwise coarse extraction goes out
Profile is used as left ventricle profile.
The specific steps of the Morphological scale-space include：
Barycenter is found out in the left ventricle profile that coarse extraction goes out to all minimums of marginal point；
If apart from the position of maximum Ma between two adjacent minimum points, using the two minimum points as segmentation
Point makees straight line and separates left ventricle profile and sustainer profile；
Redefine the barycenter of left ventricle profile；
The distance of two cut-points to barycenter is taken to put down in be connected with the barycenter center line direction of angulation of two cut-points
Average makees the 3rd cut-point；
Make curve completion left ventricle contour edge with these three cut-points, obtain left ventricle profile.
Beneficial effect：
The cardiac left-ventricle image segmenting system and method for the present invention can exclude left ventricle and active pulse-phase in terms of image
Influence even, accurate Ground Split left ventricle bottom layer image.Judge whether left ventricle is connected with sustainer and whether causes edge to be let out
The situation of dew, if producing the situation of edge leakage, morphology removes sustainer edge and uses the edge of missing and left ventricle
The similar curve in edge makes up, and overcomes the influence of sustainer edge leakage caused by left ventricle, so as to obtain the accurate left heart
Room segmentation result.
Brief description of the drawings
Fig. 1 is the cardiac left-ventricle image segmenting system block diagram of the embodiment of the present invention 1；
Fig. 2 is the cardiac left-ventricle image dividing method flow chart of the embodiment of the present invention 2；
Fig. 3 is the flow chart of the essence extraction left ventricle profile of the embodiment of the present invention 2；
Fig. 4 is the flow chart of the Morphological scale-space of the embodiment of the present invention 2；
Fig. 5 is the binary image of the embodiment of the present invention 2；
Fig. 6 is the conversion results of the Euclidean Distance Transform of the embodiment of the present invention 2；
Fig. 7 is the left ventricle profile coarse extraction result figure of the embodiment of the present invention 2；
Fig. 8 is the left ventricle profile barycenter of the embodiment of the present invention 2 to each marginal point distance Curve figure；
Fig. 9 is one group of left ventricle bottom (Ma-Mi)/MA ratio curve figures of the embodiment of the present invention 2；
Figure 10 is that the embodiment of the present invention 2 with three cut-points makees curve completion left ventricle contour edge result figure；
Figure 11 is the left ventricle profile results figure that the essence extraction of the embodiment of the present invention 2 obtains.
Embodiment
Elaborate with reference to the accompanying drawings and examples to the embodiment of the present invention.
Embodiment 1
A kind of cardiac left-ventricle image segmenting system, as shown in Figure 1, including：
Image conversion unit：The selected seed point on the left ventricle of nuclear magnetic resonance image, obtains nuclear magnetic resonance image most
High-gray level value Z0 and minimum gradation value Z1, using the average value of the maximum gradation value of nuclear magnetic resonance image and minimum gradation value as just
Beginning segmentation threshold T=(Z0+Z1)/2, is divided into foreground and background by nuclear magnetic resonance image, obtains the average gray value of prospect respectively
The average gray value T1 of T0 and background, the average gray value of the average gray value of prospect and background are taken average as new segmentation
Threshold value, if the difference of new segmentation threshold TT=(T0+T1)/2 and the segmentation threshold obtained by previous iteration is less than the permission model of setting
Enclose, then current segmentation threshold is final segmentation threshold, otherwise continues to iterate to calculate, finally obtains binary image, then pass through
Binary image is converted into gray level image by Euclidean Distance Transform；Image conversion unit extracts main device in nuclear magnetic resonance image
Official and tissue, filter out useless region, and complicated gray level image is converted into regional center gray value minimum and is passed to edge
The simple gray-scale image of increasing.
Left ventricle profile coarse extraction unit：Sort by pixel of the incremental order of gray scale to gray level image, use one
FIFO queues recursively distribute to each gray scale very small region in the way of breadth-first and form the minimum area of new gray scale respectively
Domain, and the pixel of first gray scale very small region is collectively labeled as 0 in distribution, the mark of gray scale very small region below according to
It is secondary to add 1, each linking area in gray level image is distinguished by mark, left ventricle is gone out by the seed point coarse extraction of selection
Profile.
Left ventricle profile essence extraction unit：If the left ventricle profile that coarse extraction goes out is connected with sustainer profile, master is removed
Artery, repairs left ventricle profile, obtains cardiac left-ventricle image segmentation result；If left ventricle profile and sustainer that coarse extraction goes out
Profile is not attached to, then the left ventricle profile that coarse extraction goes out is used as left ventricle profile, obtains cardiac left-ventricle image segmentation result.
Left ventricle profile essence extraction unit, including：
Analyze judging unit：Empirical value, meter are obtained by 400 groups of heart left ventricle's bottom datas for handling 10 patients
The left ventricle profile barycenter that goes out of coarse extraction is calculated to the distance of each marginal point, is found out apart from maximum Ma and apart from minimum M i, and
According to analyzing apart from maximum and apart from minimum value and judge whether left ventricle is connected with sustainer：If Ma<30 pixels
Unit length and (Ma-Mi)/Ma>0.55, or Ma>30 pixel unit lengths and (Ma-Mi)/Ma>0.38, then left ventricle
It is connected with sustainer, it is necessary to carry out Morphological scale-space, the left ventricle profile that otherwise coarse extraction goes out is used as left ventricle profile.
Morphological scale-space unit：Barycenter is found out in the left ventricle profile that coarse extraction goes out to all satisfaction (Pi- of marginal point
Mi)/(Ma-Mi)<0.3 minimum point Pi, if apart from the position of maximum Ma between two adjacent minimum points, with
The two minimum points make straight line for cut-point and separate left ventricle profile and sustainer profile, redefine left ventricle profile
Barycenter, take two cut-points to the range averaging of barycenter in be connected with the barycenter center line direction of angulation of two cut-points
Value makees the 3rd cut-point, makees curve completion left ventricle contour edge with these three cut-points, obtains left ventricle profile.
Using system provided by the invention, the cut-point of left ventricle and sustainer can be more accurately found, so as to keep away
The influence that sustainer splits left ventricle is exempted from, accurate left ventricle segmentation knot is provided for the three-dimensional reconstruction of heart left ventricle
Fruit so that the observation to left ventricle is more accurate and effective.
Embodiment 2
The present invention also provides a kind of method that cardiac left-ventricle image segmentation is carried out using system described in embodiment 1, such as Fig. 2
It is shown, including：
Step 201, the selected seed point on the left ventricle of nuclear magnetic resonance image, obtain the maximum gray scale of nuclear magnetic resonance image
Value Z0 and minimum gradation value Z1, using the average value of the maximum gradation value of nuclear magnetic resonance image and minimum gradation value as initial segmentation
Threshold value T=(Z0+Z1)/2, foreground and background is divided into by nuclear magnetic resonance image, obtains the average gray value T0 and the back of the body of prospect respectively
The average gray value T1 of scape, the average gray value of the average gray value of prospect and background is taken it is average as new segmentation threshold,
If the difference of the segmentation threshold obtained by new segmentation threshold TT=(T0+T1)/2 and previous iteration is less than the allowed band of setting, when
Preceding segmentation threshold is final segmentation threshold, otherwise continues to iterate to calculate, and finally obtains binary image as shown in Figure 5, then
Binary image is converted into by Euclidean Distance Transform by gray level image as shown in Figure 6；Image conversion unit extracts nuclear-magnetism
Major organs and tissue in resonance image, filter out useless region, and complicated gray level image is converted into regional center gray scale
Value is minimum and to the incremental simple gray-scale image in edge.
Step 202, by the incremental order of gray scale to gray level image pixel sort, using a FIFO queue according to width
Spend preferential mode and recursively distribute to each gray scale very small region and form new gray scale very small region respectively, and will in distribution
The pixel of first gray scale very small region is collectively labeled as 0, and the mark of gray scale very small region below adds 1 successively, will by mark
Each linking area in gray level image distinguishes, and left ventricle wheel as shown in Figure 7 is gone out by the seed point coarse extraction of selection
It is wide；
Step 203, essence extraction left ventricle profile：If the left ventricle profile that coarse extraction goes out is connected with sustainer profile, turn
Go step 204；If the left ventricle profile that coarse extraction goes out is not attached to sustainer profile, the left ventricle profile that coarse extraction goes out is made
For left ventricle profile, turn to go step 205；
Step 204, remove sustainer, repairs left ventricle profile；
Step 205, obtain cardiac left-ventricle image segmentation result.
As shown in figure 3, essence extraction left ventricle profile comprises the following steps that：
The left ventricle profile barycenter that step 301, calculating coarse extraction go out obtains as shown in Figure 8 to the distance of each marginal point
Curve map；
Step 302, find out apart from maximum Ma and apart from minimum M i；
If step 303, Ma<30 pixel unit lengths and (Ma-Mi)/Ma>0.55, then turn step 305, otherwise
Turn to go step 304；
If step 304, Ma>30 pixel unit lengths and (Ma-Mi)/Ma>0.38, then turn step 305, otherwise
Turn to go step 306；
Step 305, left ventricle are connected with sustainer, and left ventricle profile is obtained after carrying out Morphological scale-space；
The left ventricle profile that step 306, coarse extraction go out is used as left ventricle profile.
Fig. 9 is one group of left ventricle bottom (Ma-Mi)/MA ratio curve figures, Ma in this group of data>30 pixel unit length
Degree, (Ma-Mi)/Ma is all higher than 0.38 as seen from the figure, all there is a situation where that left ventricle is connected with sustainer.
As shown in figure 4, Morphological scale-space comprises the following steps that：
Step 401, find out in the left ventricle profile that coarse extraction goes out barycenter to all minimums of marginal point；
If step 402, apart from the position of maximum Ma between two adjacent minimum points, turn step 403, it is no
Then return to step 401；
Step 403, make straight line as cut-point using the two minimum points and separate left ventricle profile and sustainer profile；
Step 404, the barycenter for redefining left ventricle profile；
Step 405, in be connected with the barycenter center line direction of angulation of two cut-points take two cut-points to barycenter
Distance average make the 3rd cut-point；
Step 406, make curve completion left ventricle contour edge with these three cut-points, as shown in Figure 10.Pass through selection
Seed point takes connected region, extracts left ventricle profile, obtains more accurately left ventricle profile as shown in figure 11.
As it can be seen that method provided by the invention can not only separate left ventricle bottom with sustainer, additionally it is possible to approximation
The curve of left ventricle profile makes up the left ventricle profile of missing, so as to obtain accurate left ventricle profile.
Claims (7)
- A kind of 1. cardiac left-ventricle image segmenting system, it is characterised in that including：Image conversion unit：The selected seed point on the left ventricle of nuclear magnetic resonance image, by the maximum gray scale of nuclear magnetic resonance image Nuclear magnetic resonance image is divided into foreground and background, by prospect by the average value of value and minimum gradation value as initial segmentation threshold value Average gray value and background average gray value take it is average as new segmentation threshold, if new segmentation threshold and previous iteration The difference of the segmentation threshold of gained is less than the allowed band of setting, then current segmentation threshold is final segmentation threshold, is otherwise continued Iterative calculation, finally obtains binary image, then binary image is converted into gray level image by Euclidean Distance Transform；Left ventricle profile coarse extraction unit：Sort by pixel of the incremental order of gray scale to gray level image, use a FIFO team Row are recursively distributed to each center gray value minimum in gray level image in the way of breadth-first and are incremented by edge Gray areas to form new center gray value respectively minimum and to the incremental gray areas in edge, and in distribution by first Regional center gray value is minimum and is collectively labeled as 0 to the pixel of the incremental gray areas in edge, and center gray value below is most Small and to the incremental gray areas in edge mark adds 1 successively, distinguishes each linking area in gray level image by mark Open, left ventricle profile is gone out by the seed point coarse extraction of selection；Left ventricle profile essence extraction unit：If the left ventricle profile that coarse extraction goes out is connected with sustainer profile, sustainer is removed, Left ventricle profile is repaired, obtains cardiac left-ventricle image segmentation result；If the left ventricle profile that coarse extraction goes out and sustainer profile It is not attached to, then the left ventricle profile that coarse extraction goes out is used as left ventricle profile, obtains cardiac left-ventricle image segmentation result.
- 2. cardiac left-ventricle image segmenting system according to claim 1, it is characterised in that the left ventricle profile essence carries Unit is taken, including：Analyze judging unit：The left ventricle profile barycenter that calculating coarse extraction goes out is found out apart from maximum to the distance of each marginal point Ma and apart from minimum M i, and according to apart from maximum and analyzed apart from minimum value and judge left ventricle whether with active pulse-phase Even：If Ma<30 pixel unit lengths and (Ma-Mi)/Ma>0.55, or Ma>30 pixel unit lengths and (Ma- Mi)/Ma>0.38, then left ventricle be connected with sustainer, it is necessary to carry out Morphological scale-space, the left ventricle profile that otherwise coarse extraction goes out It is used as left ventricle profile.
- 3. cardiac left-ventricle image segmenting system according to claim 2, it is characterised in that the left ventricle profile essence carries Unit is taken, is further included：Morphological scale-space unit：Barycenter in the left ventricle profile that coarse extraction goes out is found out to all minimums of marginal point, if distance Then make straight line by the left heart by cut-point of the two minimum points between two adjacent minimum points in the position of maximum Ma Room profile is separated with sustainer profile, redefines the barycenter of left ventricle profile, is connected in two cut-points with the barycenter institute Angled center line direction takes the distance average of two cut-points to barycenter to make the 3rd cut-point, is made with these three cut-points Curve completion left ventricle contour edge, obtains left ventricle profile.
- 4. cardiac left-ventricle image segmenting system according to claim 3, it is characterised in that the minimum point Pi meets (Pi-Mi)/(Ma-Mi)<0.3。
- A kind of 5. cardiac left-ventricle image dividing method, it is characterised in that including：Image is changed：The selected seed point on the left ventricle of nuclear magnetic resonance image, by the maximum gradation value of nuclear magnetic resonance image and Nuclear magnetic resonance image is divided into foreground and background, by the flat of prospect by the average value of minimum gradation value as initial segmentation threshold value Gray value and the average gray value of background take average as new segmentation threshold, if new segmentation threshold and previous iteration gained Segmentation threshold difference be less than setting allowed band, then current segmentation threshold is final segmentation threshold, otherwise continues iteration Calculate, finally obtain binary image, then binary image is converted into by gray level image by Euclidean Distance Transform；Coarse extraction left ventricle profile：Sort by pixel of the incremental order of gray scale to gray level image, pressed using a fifo queue According to the mode of breadth-first, recursively to distribute to each center gray value in gray level image minimum and to the incremental gray scale in edge Region forms new center gray value minimum and to the incremental gray areas in edge respectively, and in distribution by first center ash Angle value is minimum to be simultaneously collectively labeled as 0 to the pixel of the incremental gray areas in edge, and center gray value below is minimum and to edge The mark of incremental gray areas adds 1 successively, distinguishes each linking area in gray level image by mark, passes through selection Seed point coarse extraction go out left ventricle profile；Essence extraction left ventricle profile：If the left ventricle profile that coarse extraction goes out is connected with sustainer profile, sustainer is removed, is repaired Left ventricle profile, obtains cardiac left-ventricle image segmentation result；If the left ventricle profile that coarse extraction goes out and sustainer profile not phase Even, then the left ventricle profile that coarse extraction goes out is used as left ventricle profile, obtains cardiac left-ventricle image segmentation result.
- 6. according to the method described in claim 5, it is characterized in that, the specific steps of the essence extraction left ventricle profile include：Distance of the left ventricle profile barycenter that calculating coarse extraction goes out to each marginal point；Find out apart from maximum Ma and apart from minimum M i；According to analyzing apart from maximum and apart from minimum value and judge whether left ventricle is connected with sustainer：If Ma<30 pictures Vegetarian refreshments unit length and (Ma-Mi)/Ma>0.55, or Ma>30 pixel unit lengths and (Ma-Mi)/Ma>0.38, then it is left Ventricle obtains left ventricle profile after being connected with sustainer, it is necessary to carry out Morphological scale-space, the left ventricle profile that otherwise coarse extraction goes out It is used as left ventricle profile.
- 7. according to the method described in claim 6, it is characterized in that, the specific steps of the Morphological scale-space include：Barycenter is found out in the left ventricle profile that coarse extraction goes out to all minimums of marginal point；If apart from the position of maximum Ma between two adjacent minimum points, make by cut-point of the two minimum points Straight line separates left ventricle profile and sustainer profile；Redefine the barycenter of left ventricle profile；Two cut-points are taken to the distance average of barycenter in be connected with the barycenter center line direction of angulation of two cut-points Make the 3rd cut-point；Make curve completion left ventricle contour edge with these three cut-points, obtain left ventricle profile.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610161395.2A CN105741310B (en) | 2016-03-21 | 2016-03-21 | A kind of cardiac left-ventricle image segmenting system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610161395.2A CN105741310B (en) | 2016-03-21 | 2016-03-21 | A kind of cardiac left-ventricle image segmenting system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105741310A CN105741310A (en) | 2016-07-06 |
| CN105741310B true CN105741310B (en) | 2018-05-04 |
Family
ID=56251733
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610161395.2A Active CN105741310B (en) | 2016-03-21 | 2016-03-21 | A kind of cardiac left-ventricle image segmenting system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105741310B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107220984B (en) * | 2017-05-05 | 2021-07-16 | 上海联影医疗科技股份有限公司 | Image segmentation method, image segmentation system and image segmentation device |
| CN107610095A (en) * | 2017-08-04 | 2018-01-19 | 南京邮电大学 | Heart CT coronary artery full-automatic partition methods based on image co-registration |
| CN109272512B (en) * | 2018-09-25 | 2022-02-15 | 南昌航空大学 | Method for automatically segmenting left ventricle inner and outer membranes |
| 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 |
| CN110378948B (en) * | 2019-07-02 | 2023-08-25 | 北京字节跳动网络技术有限公司 | 3D model reconstruction method and device and electronic equipment |
| CN110471469B (en) * | 2019-09-26 | 2021-03-16 | 湖南航天机电设备与特种材料研究所 | Self-adaptive temperature control method based on PI algorithm |
| CN111028254B (en) * | 2019-12-09 | 2023-03-24 | 上海交通大学医学院附属上海儿童医学中心 | Left myocardium segmentation method based on cardiac cine magnetic resonance image |
| CN111369525B (en) * | 2020-03-02 | 2023-05-30 | 联影智能医疗科技(北京)有限公司 | Image analysis method, apparatus and storage medium |
| CN111815586B (en) * | 2020-06-29 | 2022-08-05 | 苏州润迈德医疗科技有限公司 | Method and system for acquiring connected domain of left atrium and left ventricle based on CT image |
| CN112215839B (en) * | 2020-10-27 | 2024-04-19 | 东软教育科技集团有限公司 | Six-segment segmentation method, system and storage medium for left ventricular wall of echocardiogram |
| CN112308845B (en) * | 2020-11-03 | 2021-07-02 | 赛诺威盛科技(北京)股份有限公司 | Left ventricle segmentation method and device and electronic equipment |
| CN112330708B (en) * | 2020-11-24 | 2024-04-23 | 沈阳东软智能医疗科技研究院有限公司 | Image processing method, device, storage medium and electronic equipment |
| CN112614062B (en) * | 2020-12-17 | 2024-04-26 | 北京陆桥技术股份有限公司 | Colony counting method, colony counting device and computer storage medium |
| CN112712530A (en) * | 2020-12-24 | 2021-04-27 | 中国人民解放军陆军军医大学第一附属医院 | Cardiac MRI left ventricle segmentation method and device |
| CN113160116B (en) * | 2021-02-03 | 2022-12-27 | 中南民族大学 | Method, system and equipment for automatically segmenting inner membrane and outer membrane of left ventricle |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102397070A (en) * | 2011-01-26 | 2012-04-04 | 无锡宙示医学软件开发有限公司 | Method for full-automatic segmentation and quantification of left ventricle of cardiac magnetic resonance image |
| CN103886603A (en) * | 2014-03-31 | 2014-06-25 | 西北工业大学 | Left ventricle nuclear magnetic resonance image segmentation and three-dimensional reconstruction method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5597492B2 (en) * | 2010-09-08 | 2014-10-01 | 株式会社東芝 | Ultrasonic diagnostic apparatus, image processing apparatus, and program |
-
2016
- 2016-03-21 CN CN201610161395.2A patent/CN105741310B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102397070A (en) * | 2011-01-26 | 2012-04-04 | 无锡宙示医学软件开发有限公司 | Method for full-automatic segmentation and quantification of left ventricle of cardiac magnetic resonance image |
| CN103886603A (en) * | 2014-03-31 | 2014-06-25 | 西北工业大学 | Left ventricle nuclear magnetic resonance image segmentation and three-dimensional reconstruction method |
Non-Patent Citations (2)
| Title |
|---|
| 基于活动轮廓模型的左心室MR图像分割;张建伟 等;《电子学报》;20111130(第11期);第2670-2673页 * |
| 结合分水岭和区域合并的彩色图像分割;李小红 等;《电子测量与仪器学报》;20130331;第27卷(第3期);第247-252页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105741310A (en) | 2016-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105741310B (en) | A kind of cardiac left-ventricle image segmenting system and method | |
| CN105957066B (en) | CT image liver segmentation method and system based on automatic context model | |
| CN104933709B (en) | Random walk CT lung tissue image automatic segmentation methods based on prior information | |
| CN102385751B (en) | Liver tumor region segmentation method based on watershed transform and classification through support vector machine | |
| CN103886599B (en) | A kind of blood vessel ROI dividing methods based on ivus image | |
| CN104143184B (en) | A kind of method of lung segmentation | |
| CN103473767B (en) | The method and system that a kind of soft tissues of abdomen nuclear-magnetism image is cut apart | |
| CN106780497B (en) | A kind of organ vascular tree extraction method based on statistical information | |
| JP2023517058A (en) | Automatic detection of tumors based on image processing | |
| Hadhoud et al. | Left ventricle segmentation in cardiac MRI images | |
| CN108230301A (en) | A kind of spine CT image automatic positioning dividing method based on active contour model | |
| CN107680110B (en) | Inner ear three-dimensional level set segmentation method based on statistical shape model | |
| CN103035009A (en) | Pulmonary nodule edge rebuilding and partitioning method based on computed tomography (CT) image | |
| Zidan et al. | Level set-based CT liver image segmentation with watershed and artificial neural networks | |
| CN110349168B (en) | Femoral head CT image segmentation method | |
| CN105389821B (en) | It is a kind of that the medical image cutting method being combined is cut based on cloud model and figure | |
| Ninh et al. | Skin lesion segmentation based on modification of SegNet neural networks | |
| CN106846330A (en) | Human liver's feature modeling and vascular pattern space normalizing method | |
| Gao et al. | Accurate lung segmentation for X-ray CT images | |
| CN106780491B (en) | Initial contour generation method adopted in segmentation of CT pelvic image by GVF method | |
| Xia et al. | 3D cascaded convolutional networks for multi-vertebrae segmentation | |
| Sagar et al. | Color channel based segmentation of skin lesion from clinical images for the detection of melanoma | |
| CN108961278B (en) | Method and system for abdominal wall muscle segmentation based on image data | |
| CN108280833B (en) | Skeleton extraction method for plant root system bifurcation characteristics | |
| CN104933723B (en) | Tongue image dividing method based on rarefaction representation |
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 |