CN105243666A - Medical MR image segmentation method based on Hough transform and geometric active contour - Google Patents
Medical MR image segmentation method based on Hough transform and geometric active contour Download PDFInfo
- Publication number
- CN105243666A CN105243666A CN201510654070.3A CN201510654070A CN105243666A CN 105243666 A CN105243666 A CN 105243666A CN 201510654070 A CN201510654070 A CN 201510654070A CN 105243666 A CN105243666 A CN 105243666A
- Authority
- CN
- China
- Prior art keywords
- left ventricle
- hough transform
- profile
- segmentation
- outside contour
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- 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
- G06T2207/20048—Transform domain processing
- G06T2207/20061—Hough transform
-
- 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)
- Quality & Reliability (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The present invention discloses a medical MR image segmentation method based on Hough transform and a geometric active contour and belongs to the technical field of image processing. According to the method, firstly, the prior shape knowledge that inner and outer contours of a left ventricle myocardium on a short axis image are of a round-like shaped is utilized, Hough transform is adopted to estimate an initial contour of a left ventricle and the reason of adopting Hough transform is that Hough transform has strong robustness, is not quite sensitive for the incompleteness of data or noise and can recognize partially deformed or partially shielded objects; the reason of using a K-means clustering algorithm is that as a square error based clustering method, the K-means clustering algorithm is simple and has a high clustering speed; and by using the method provided by the present invention, the inner and outer contours of the left ventricle can be segmented effectively, relative positions of evolving curves of the inner and outer contours of the left ventricle can be controlled and a function of shape constraint can be realized.
Description
Technical field:
The present invention relates to image to distort technical field, be specifically related to the medical science MR image partition method based on Hough transform and geometry initiatively profile.
Background technology:
For the segmentation of myocardium of left ventricle, main stream approach is the image partition method driven based on edge.As: parameter active contour model (also known as Snake model), level set (LevelSet) model, geometric active contour model and their improved model.But usually there is the phenomenons such as papillary muscle interference and partial gradient maximum value region, weak edge, artifact in MR image, bring difficulty to the image partition method based on edge.For the problem that these phenomenons are brought, many prioris Image Segmentation Model be introduced into based on edge improves the robustness of model.More be shape constraining based on priori and area information, this algorithm, on the basis of geometric active contour model, has considered the edge of image and area information and physiological structure constraint, can split the inside and outside contour of left ventricle simultaneously.
But the method for Paragios also exists some problems, they adopt GVF (GradientVectorFlow) to replace gradient fields guiding curve to develop to edge, can expand the catching range of model like this.But for image, noisy image that gray scale is uneven, the performance of GVF will be affected, or not very good GVF field can only be obtained, or the structure of target may be floating.Now evolution curve will enter another region from weak boundary, causes boundary leaking phenomenon or stops developing at the local maximum place of intra-zone gradient and isolated edge, can not move to real border.Although the method for Paragios adds the area information obtained by gauss hybrid models matching histogram, but this method is better to having three more satisfactory peak image fitting effect, and the EM algorithm being used for solving gauss hybrid models has certain dependence for initial parameter estimation, may be stuck in certain local maximum.In addition, the method for Paragios due to the item number added more, curve evolvement formula is more multiple, so in order to ensure the numerical computation method that its numerical stability sexual needs are special, make numerical evaluation complexity higher like this.
Summary of the invention:
For the problems referred to above, the technical problem to be solved in the present invention be to provide a kind of effective, analyze accurately based on the medical science MR image partition method of Hough transform and geometry active profile.
A kind of medical science MR image partition method based on Hough transform and geometry initiatively profile of the present invention, it comprise following some:
1, the prior shape knowledge of the inside and outside contour sub-circular of myocardium of left ventricle in short axis images is utilized, the initial profile of left ventricle is first automatically located by Hough transform, initial profile is made to be positioned actual profile adjacent edges more exactly, and then on the basis of geometric active contour model, utilize and by K-mean cluster, the physiological structure that the target in image carries out area information that coarse segmentation provides and cardiac muscle is retrained, set up the myocardium inside and outside contour curve evolvement equation that lotus root is closed, auto Segmentation is carried out to the inside and outside contour of left ventricle simultaneously;
2, K-means clustering algorithm
Cluster is exactly, by certain similarity measurement criterion, Data Placement is become the many groups of subclasses with similar properties, makes the similarity of subclass inside be greater than similarity between class; Therefore the tolerance of distance or similarity is the basis of clustering algorithm.
K-mean cluster belongs to unsupervised learning method, the application in Iamge Segmentation, is generally target in image is divided into different classes by gray-scale value, or combining image histogram carries out cluster segmentation image.In the feature space of image slices vegetarian refreshments, carry out cluster is exactly the pixel group finding feature similarity in this space.According to the difference of initial cluster center system of selection, the difference of sample allocation rule, the difference of centroid computing method, the difference of stopping criterion, has the algorithm of multiple different editions.K-mean algorithm is when cluster initial, and cluster centre is generally random generation, and cluster number is also manual setting.For the selection of initial cluster center, it can have a huge impact final result, generally has front several method: (1) selects special sample; (2) K the sample started is selected; (3) all samples are arranged by distance, select the sample of few for these range differences decile; (4) select to make between class distance maximum sample;
3, the location of the initial internal and external contour line of left ventricle
It can according to the priori of approximate object model, detects and does not know the object of accurate shape; Utilize the prior shape knowledge of the similar circle of the inside and outside contour of myocardium of left ventricle in short axis images, adopt Hough transform to estimate left ventricle inside and outside contour; In left ventricular contraction process, when left ventricle Internal periphery due to papillary muscle impact no longer sub-circular time, there will be the phenomenon that can't detect left ventricle Internal periphery, so detect the outline of left ventricle by Hough transform, the center of circle of recycling outline combines the initial left ventricle profile of Internal periphery radius estimated by MR imaging parameters; Left ventricle Internal periphery marginal information and area grayscale consistance are all good than outline, so the dependence of the segmentation of left ventricle Internal periphery to initial profile is much smaller.
Hough transform can identification division distortion or the object of partial occlusion, but its calculated amount is also larger.In order to reduce the complexity of calculating, improving the accuracy of algorithm, the intersection point one of long axis and short axis plane can be utilized to fix on ventricle inside, and near left ventricle inside and outside contour central point this priori.
The present invention first utilizes the prior shape knowledge of the similar circle of the inside and outside contour of myocardium of left ventricle in short axis images, Hough transform is adopted to estimate the initial profile of left ventricle, initial profile is made to be positioned actual profile adjacent edges more exactly, and then on the basis of geometry active contour model, K-mean cluster is utilized to carry out the physiological structure constraint of area information that coarse segmentation provides and cardiac muscle to the target in image, set up the myocardium inside and outside contour curve evolvement equation of coupling, automatic segmentation algorithm is carried out to the inside and outside contour of left ventricle simultaneously.
Use the reason of Hough transform to be its strong robustness, to data not exclusively or noise be not very responsive, can the object of identification division distortion or partial occlusion.Estimate initial inside and outside contour by such method, initial profile can be made to be positioned actual profile adjacent edges more exactly, the curve evolvement be convenient to below converges to correct profile, can overcome the impact of initial profile on geometry active contour model segmentation result.For the segmentation of left ventricle inside and outside contour, the simple Image Segmentation Model based on edge often can not get desirable segmentation effect, so many prioris are introduced into model to improve the robustness of model.More be shape constraining based on priori and target area information, the physiological structure constraint of cardiac muscle can be used for controlling the relative position of left ventricle inside and outside contour, and area information can be used for providing target classification information to Iamge Segmentation.
Use K-means clustering algorithm to be because it is as a kind of clustering method based on square error, algorithm is simple, and cluster speed is fast.And the inconsistent and histogram of MR image-region gray scale is not form desirable three peak Distribution by left ventricle inner membrance with inner region, myocardial region, background area, therefore, K-mean algorithm classifying quality is used to be better than gauss hybrid models.
Beneficial effect of the present invention: it splits left ventricle inside and outside contour effectively, the segmentation result of additive method then departs from actual profile.And by comparing the effect that it can also be seen that each add-ins: Internal periphery area information item; When curve evolvement is inner to left ventricle Internal periphery, curve will expand; Time beyond curve evolvement to Internal periphery, curve will shrink.This role is equivalent to an adaptive Balloon Force.The physiological structure bound term of cardiac muscle; Control the relative position of left ventricle inside and outside contour evolution curve, play a part shape constraining.
Embodiment:
This embodiment is by the following technical solutions: it comprise following some:
1, the prior shape knowledge of the inside and outside contour sub-circular of myocardium of left ventricle in short axis images is utilized, the initial profile of left ventricle is first automatically located by Hough transform, initial profile is made to be positioned actual profile adjacent edges more exactly, and then on the basis of geometric active contour model, utilize and by K-mean cluster, the physiological structure that the target in image carries out area information that coarse segmentation provides and cardiac muscle is retrained, set up the myocardium inside and outside contour curve evolvement equation that lotus root is closed, auto Segmentation is carried out to the inside and outside contour of left ventricle simultaneously.
2, K-means clustering algorithm
Cluster is exactly, by certain similarity measurement criterion, Data Placement is become the many groups of subclasses with similar properties, makes the similarity of subclass inside be greater than similarity between class.Therefore the tolerance of distance or similarity is the basis of clustering algorithm.
K-means clustering algorithm is a kind of algorithm of continuous iteration adjustment k the most general cluster barycenter.It is a kind of clustering method based on square error, is also a hard clustering algorithm very famous in clustering algorithm, and this algorithm is simple, and cluster speed is fast.The same with other clustering algorithm, K-mean cluster is also the process of an iteration optimizing.
Supposing will by sample set data X={x
1, x
2..., x
nbe divided into K class, first select the initial division of a K class, calculate the mean vector μ of these classes, then according to Euclidean distance, remaining each sample is assigned to its nearest division of class mean distance.
Recalculate the mean vector of the sample being assigned to each class, as Xin Lei center.Repeat this process until mean vector convergence.
The central idea of K-means clustering algorithm minimizes total inter-object distance.All methods repeat to adjust the center of each cluster of k, and be assigned to by each sample in the classification at nearest barycenter place.Following formula gives total class distance of euclidean distance metric:
The concrete steps of algorithm are as follows:
Step1: the minimum deflection threshold value of setting maximum iteration time and permission, the center of all classes of initialization
Step2: x
ibe assigned to it at a distance of that nearest barycenter
in representative cluster;
Step3: the new barycenter after dispensed
and E
(n+1);
Step4: repeat Step2 and Step3, until || E
(n+1)-E
(n)|| stop when being less than threshold value or reaching maximum iteration time.
K-mean cluster belongs to unsupervised learning method, the application in Iamge Segmentation, is generally target in image is divided into different classes by gray-scale value, or combining image histogram carries out cluster segmentation image.In the feature space of image slices vegetarian refreshments, carry out cluster is exactly the pixel group finding feature similarity in this space.According to the difference of initial cluster center system of selection, the difference of sample allocation rule, the difference of centroid computing method, the difference of stopping criterion, has the algorithm of multiple different editions.K-mean algorithm is when cluster initial, and cluster centre is generally random generation, and cluster number is also manual setting.For the selection of initial cluster center, it can have a huge impact final result, generally has front several method: (1) selects special sample; (2) K the sample started is selected; (3) all samples are arranged by distance, select the sample of few for these range differences decile; (4) select to make between class distance maximum sample
3, the location of the initial internal and external contour line of left ventricle
In order to overcome the impact of initial profile on geometry active contour model segmentation result, Hough transform is adopted to estimate initial internal and external contour line.Hough transform is the method for the parsing definition shape existed in conventional detected image.It can according to the priori of approximate object model, detects and does not know the object of accurate shape.Utilize the prior shape knowledge of the similar circle of the inside and outside contour of myocardium of left ventricle in short axis images, adopt Hough transform to estimate left ventricle inside and outside contour.In left ventricular contraction process, when left ventricle Internal periphery due to papillary muscle impact no longer sub-circular time, there will be the phenomenon that can't detect left ventricle Internal periphery, so detect the outline of left ventricle by Hough transform, the center of circle of recycling outline combines the initial left ventricle profile of Internal periphery radius estimated by MR imaging parameters.Left ventricle Internal periphery marginal information and area grayscale consistance are all good than outline, so the dependence of the segmentation of left ventricle Internal periphery to initial profile is much smaller, so accurately " initial outline then will be positioned actual profile adjacent edges more exactly, and the curve evolvement be convenient to below converges to correct profile not need picture outline.
Hough transform can identification division distortion or the object of partial occlusion, but its calculated amount is also larger.In order to reduce the complexity of calculating, improving the accuracy of algorithm, the intersection point one of long axis and short axis plane can be utilized to fix on ventricle inside, and near left ventricle inside and outside contour central point this priori.Centered by this intersection point, get the window (size is determined by MR imaging parameters, can obtain from MR image Dicom file header) of certain limit, in this window, calculate Hough transform.
The transformation for mula that Hough transform finds bowlder is.Initial profile line location algorithm herein:
Step1: the scope estimating left ventricle outline radius according to MR imaging parameters;
Step2: allow the radius of Hough transform value in value range, the number of the circle that Hough transform detects under record different radii;
Step3: get and make as minimizing radius is as left ventricle outline radius (if the center mean value of the he of close long axis and short axis plane intersection point is got in the center of circle);
Step4: the initial left ventricle profile of Internal periphery radius utilizing the center of circle of outline to combine to be estimated by MR imaging parameters.
Left ventricle inside and outside contour is split in the present invention effectively, and the segmentation result of additive method then departs from actual profile.And by comparing the effect that it can also be seen that each add-ins: Internal periphery area information item; When curve evolvement is inner to left ventricle Internal periphery, curve will expand; Time beyond curve evolvement to Internal periphery, curve will shrink.This role is equivalent to an adaptive Balloon Force.The physiological structure bound term of cardiac muscle; Control the relative position of left ventricle inside and outside contour evolution curve, play a part shape constraining.
More than show and describe ultimate principle of the present invention and principal character and advantage of the present invention.The technician of the industry should understand; the present invention is not restricted to the described embodiments; what describe in above-described embodiment and instructions just illustrates principle of the present invention; without departing from the spirit and scope of the present invention; the present invention also has various changes and modifications, and these changes and improvements all fall in the claimed scope of the invention.Application claims protection domain is defined by appending claims and equivalent thereof.
Claims (2)
1. based on the medical science MR image partition method of Hough transform and geometry initiatively profile, it is characterized in that it comprise following some:
(1) the prior shape knowledge of the inside and outside contour sub-circular of myocardium of left ventricle in short axis images, is utilized, the initial profile of left ventricle is first automatically located by Hough transform, initial profile is made to be positioned actual profile adjacent edges more exactly, and then on the basis of geometric active contour model, utilize and by K-mean cluster, the physiological structure that the target in image carries out area information that coarse segmentation provides and cardiac muscle is retrained, set up the myocardium inside and outside contour curve evolvement equation that lotus root is closed, auto Segmentation is carried out to the inside and outside contour of left ventricle simultaneously;
(2), K-means clustering algorithm
Cluster is exactly, by certain similarity measurement criterion, Data Placement is become the many groups of subclasses with similar properties, makes the similarity of subclass inside be greater than similarity between class; Therefore the tolerance of distance or similarity is the basis of clustering algorithm;
(3), the location of the initial internal and external contour line of left ventricle
It can according to the priori of approximate object model, detects and does not know the object of accurate shape; Utilize the prior shape knowledge of the similar circle of the inside and outside contour of myocardium of left ventricle in short axis images, adopt Hough transform to estimate left ventricle inside and outside contour; In left ventricular contraction process, when left ventricle Internal periphery due to papillary muscle impact no longer sub-circular time, there will be the phenomenon that can't detect left ventricle Internal periphery, so detect the outline of left ventricle by Hough transform, the center of circle of recycling outline combines the initial left ventricle profile of Internal periphery radius estimated by MR imaging parameters; Left ventricle Internal periphery marginal information and area grayscale consistance are all good than outline, so the dependence of the segmentation of left ventricle Internal periphery to initial profile is much smaller.
2. the medical science MR image partition method based on Hough transform and geometry initiatively profile according to claim 1, it is characterized in that it first utilizes the prior shape knowledge of the similar circle of the inside and outside contour of myocardium of left ventricle in short axis images, Hough transform is adopted to estimate the initial profile of left ventricle, initial profile is made to be positioned actual profile adjacent edges more exactly, and then on the basis of geometry active contour model, K-mean cluster is utilized to carry out the physiological structure constraint of area information that coarse segmentation provides and cardiac muscle to the target in image, set up the myocardium inside and outside contour curve evolvement equation of coupling, automatic segmentation algorithm is carried out to the inside and outside contour of left ventricle simultaneously,
Use the reason of Hough transform to be its strong robustness, to data not exclusively or noise be not very responsive, can the object of identification division distortion or partial occlusion.Estimate initial inside and outside contour by such method, initial profile can be made to be positioned actual profile adjacent edges more exactly, the curve evolvement be convenient to below converges to correct profile, can overcome the impact of initial profile on geometry active contour model segmentation result; For the segmentation of left ventricle inside and outside contour, the simple Image Segmentation Model based on edge often can not get desirable segmentation effect, so many prioris are introduced into model to improve the robustness of model.More be shape constraining based on priori and target area information, the physiological structure constraint of cardiac muscle can be used for controlling the relative position of left ventricle inside and outside contour, and area information can be used for providing target classification information to Iamge Segmentation;
Use K-means clustering algorithm to be because it is as a kind of clustering method based on square error, algorithm is simple, and cluster speed is fast; And the inconsistent and histogram of MR image-region gray scale is not form desirable three peak Distribution by left ventricle inner membrance with inner region, myocardial region, background area, therefore, K-mean algorithm classifying quality is used to be better than gauss hybrid models.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510654070.3A CN105243666A (en) | 2015-10-09 | 2015-10-09 | Medical MR image segmentation method based on Hough transform and geometric active contour |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510654070.3A CN105243666A (en) | 2015-10-09 | 2015-10-09 | Medical MR image segmentation method based on Hough transform and geometric active contour |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105243666A true CN105243666A (en) | 2016-01-13 |
Family
ID=55041298
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510654070.3A Pending CN105243666A (en) | 2015-10-09 | 2015-10-09 | Medical MR image segmentation method based on Hough transform and geometric active contour |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105243666A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106124535A (en) * | 2016-06-12 | 2016-11-16 | 上海市建筑科学研究院 | The bar connecting of concrete cover grout sleeve compactness inspection method |
| CN107067393A (en) * | 2017-04-14 | 2017-08-18 | 大连理工大学 | A kind of three-dimensional medical image segmentation method based on user mutual and shape prior knowledge |
| CN110232695A (en) * | 2019-06-18 | 2019-09-13 | 山东师范大学 | Left ventricle image partition method and system based on hybrid mode image |
| CN110739050A (en) * | 2019-12-20 | 2020-01-31 | 深圳大学 | left ventricle full parameter and confidence degree quantification method |
| CN111028254A (en) * | 2019-12-09 | 2020-04-17 | 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 |
| CN113610810A (en) * | 2021-08-09 | 2021-11-05 | 华力创科学(深圳)有限公司 | Blood vessel detection method based on Markov random field |
| CN115631208A (en) * | 2022-10-13 | 2023-01-20 | 中国矿业大学 | Unmanned aerial vehicle image mining area ground crack extraction method based on improved active contour model |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101576378A (en) * | 2009-06-16 | 2009-11-11 | 华南理工大学 | System and method for automatically checking articles based on movable profiles |
| CN103700095A (en) * | 2013-12-10 | 2014-04-02 | 东北林业大学 | Log end surface image partitioning algorithm for improving active contour model based on circle constraint |
-
2015
- 2015-10-09 CN CN201510654070.3A patent/CN105243666A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101576378A (en) * | 2009-06-16 | 2009-11-11 | 华南理工大学 | System and method for automatically checking articles based on movable profiles |
| CN103700095A (en) * | 2013-12-10 | 2014-04-02 | 东北林业大学 | Log end surface image partitioning algorithm for improving active contour model based on circle constraint |
Non-Patent Citations (4)
| Title |
|---|
| FREDERIC GIBOU ET AL.: "A Fast Hybrid k-Means Level Set Algorithm For Segmentation", 《PROE.OF INT.CONF.ON STATISTICS AND MATHEMATICS》 * |
| 刘复昌等: "结合Hough变换与测地线轮廓模型的MR图像左心室自动分割", 《计算机辅助设计与图形学学报》 * |
| 吴暾华等: "基于FCM聚类和C-V模型的人眼精确定位方法", 《电路与系统学报》 * |
| 朱敏 等: "基于Snake改进模型的心脏MR图像左心室分割方法", 《四川大学学报（工程科学版）》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106124535A (en) * | 2016-06-12 | 2016-11-16 | 上海市建筑科学研究院 | The bar connecting of concrete cover grout sleeve compactness inspection method |
| CN107067393A (en) * | 2017-04-14 | 2017-08-18 | 大连理工大学 | A kind of three-dimensional medical image segmentation method based on user mutual and shape prior knowledge |
| CN107067393B (en) * | 2017-04-14 | 2019-08-20 | 大连理工大学 | A kind of three-dimensional medical image segmentation method based on user's interaction and shape prior knowledge |
| CN110232695A (en) * | 2019-06-18 | 2019-09-13 | 山东师范大学 | Left ventricle image partition method and system based on hybrid mode image |
| CN111028254A (en) * | 2019-12-09 | 2020-04-17 | 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 |
| 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 |
| CN110739050A (en) * | 2019-12-20 | 2020-01-31 | 深圳大学 | left ventricle full parameter and confidence degree quantification method |
| CN113610810A (en) * | 2021-08-09 | 2021-11-05 | 华力创科学(深圳)有限公司 | Blood vessel detection method based on Markov random field |
| CN115631208A (en) * | 2022-10-13 | 2023-01-20 | 中国矿业大学 | Unmanned aerial vehicle image mining area ground crack extraction method based on improved active contour model |
| CN115631208B (en) * | 2022-10-13 | 2023-06-16 | 中国矿业大学 | Unmanned aerial vehicle image mining area ground crack extraction method based on improved active contour model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105243666A (en) | Medical MR image segmentation method based on Hough transform and geometric active contour | |
| Artaechevarria et al. | Combination strategies in multi-atlas image segmentation: application to brain MR data | |
| Ding et al. | Target recognition in synthetic aperture radar images via matching of attributed scattering centers | |
| AU2018253501B2 (en) | Method and apparatus for automated delineation of structore shape for image guided treatment planning | |
| US10096108B2 (en) | Medical image segmentation method and apparatus | |
| CN104537673B (en) | Infrared Image Segmentation based on multi thresholds and adaptive fuzzy clustering | |
| Wimmer et al. | A generic probabilistic active shape model for organ segmentation | |
| US8150119B2 (en) | Method and system for left ventricle endocardium surface segmentation using constrained optimal mesh smoothing | |
| Hanek et al. | The contracting curve density algorithm: Fitting parametric curve models to images using local self-adapting separation criteria | |
| Zhang et al. | Level set evolution driven by optimized area energy term for image segmentation | |
| CN107742113B (en) | One kind being based on the posterior SAR image complex target detection method of destination number | |
| Furukawa et al. | Automatic liver segmentation method based on maximum a posterior probability estimation and level set method | |
| CN104021559A (en) | Image registration method based on mutual information and Harris corner point detection | |
| El-Baz et al. | A novel 3D joint Markov-Gibbs model for extracting blood vessels from PC–MRA images | |
| Yuan | Active contour driven by region-scalable fitting and local Bhattacharyya distance energies for ultrasound image segmentation | |
| CN112734816B (en) | Heterologous image registration method based on CSS-Delaunay | |
| CN109508674A (en) | Airborne lower view isomery image matching method based on region division | |
| Zhou et al. | Nearest neighbor field driven stochastic sampling for abrupt motion tracking | |
| CN115035326B (en) | Radar image and optical image accurate matching method | |
| Hontani et al. | Point-based non-rigid surface registration with accuracy estimation | |
| Wilhelm et al. | Boundary aware image segmentation with unsupervised mixture models | |
| Wang et al. | Shape constraints for the left ventricle segmentation from cardiac cine mri based on snake models | |
| El-Baz et al. | Robust image segmentation using learned priors | |
| Gopi et al. | Analysis of lung tumour detection and segmentation using level set method of active contour model | |
| Lv et al. | Segmenting similar shapes via weighted group-similarity active contours |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160113 |
|
| RJ01 | Rejection of invention patent application after publication |