CN109727250A - The CT image sequence pulmonary artery metamorphosis detection method indicated based on radial function - Google Patents
The CT image sequence pulmonary artery metamorphosis detection method indicated based on radial function Download PDFInfo
- Publication number
- CN109727250A CN109727250A CN201811601713.8A CN201811601713A CN109727250A CN 109727250 A CN109727250 A CN 109727250A CN 201811601713 A CN201811601713 A CN 201811601713A CN 109727250 A CN109727250 A CN 109727250A
- Authority
- CN
- China
- Prior art keywords
- pulmonary artery
- target
- radial function
- curve
- variance
- 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
Abstract
The CT image sequence pulmonary artery metamorphosis detection method indicated based on radial function belongs to technical field of medical image processing.The present invention is directed to the pulmonary artery target sequence being partitioned into, and the radial function curve of single pulmonary artery target is sought using radial function representation method, for the pulmonary artery target of entire CTPA image sequence, family of curves is just had and is corresponding to it；Variance is sought to the family of curves, obtains variance function curve；Angle corresponding to variance curve maximum value is sought, the angle direction the most significant of pulmonary artery metamorphosis in CTPA image sequence is thus obtained.
Description
Technical field
The present invention relates to a kind of changes of computed tomography pulmonary arterial vascular radiography (CTPA) image sequence pulmonary artery form
Change detection method, belongs to technical field of medical image processing.
Background technique
CTPA image pulmonary artery divide automatically be Pulmonary Arterial Diseases computer-aided diagnosis key link, for pulmonary artery
Early discovery, the early diagnosis of class disease are of great significance with early treatment.Pulmonary artery tree has complicated anatomical structure, specific manifestation
To be gradually divided into the branched structures such as leaf, section and sub- section from vessel trunk；This is resulted in CTPA image sequence, different layers
The pulmonary artery morphological differences of grade is larger, it is possible that the case where form of pulmonary artery mutates in adjacent two picture, this
Great difficulty is brought to pulmonary artery segmentation.
Target following method is a kind of basic CTPA image pulmonary artery dividing method, and active contour model, which is used as, is based on target
The relation information of the marginal information and image sequence of image or more interlayer is utilized in the target following method of edge contour, it can make
Curve convergence at target concave edges, divide by the detection suitable for this kind of irregular figure of pulmonary artery.But when in image sequence
When morphological mutation occurs for adjacent pulmonary artery, active contour model cannot accurately reach the side of pulmonary artery after target morphology mutation sometimes
Edge causes missing inspection or tracks the loss of target.Therefore, form change is carried out for the target gone out using target following method primary segmentation
The direction that target morphology is changed significantly in image sequence is determined in the detection of change, is a urgent problem to be solved.
The random hypersurface model (RHM) of star convex can model any star convex target, now be chiefly used in utilizing thunder
The tracking of the extension target carried out up to equal sensors, but it is also more rare in the application of medical image sequences target tracking domain
See.During being modeled using RHM to target, star convex need to be indicated with radial function, radial function can be analyzed to Fu
In leaf series, and the rough profile of shape can be described by the low-frequency harmonics of Fourier space, and the detailed information of shape is then by Fu
The shape that the high-frequency harmonic expression of leaf series, i.e. radial function are able to reflect target.For CTPA image pulmonary artery target,
It is feasible for describing its shape using this radial function representation method.
Summary of the invention
It is an object of the invention to solve traditional images tracking and partitioning algorithm in the segmentation of CTPA image sequence pulmonary artery
Above-mentioned technical problem, the radial function representation method of this method combining target provides a kind of CTPA image sequence pulmonary artery shape
State change detecting method.
To achieve the above object, the technical solution adopted by the present invention is a kind of target morphology change indicated based on radial function
Change detection method, this method is including the use of radial function representation method for the lung of primary segmentation out from individual each CTPA image
Artery carries out radial function description, obtains radial function curve；Cluster radial function curve is then obtained for whole image sequence
Then race seeks variance to family of curves, obtain variance curve；Angle corresponding to the maximum value of variance curve is sought later, with
Determine target sequence metamorphosis direction the most significant.
Angle for a star convex target, between the line l and trunnion axis of centroid to marginal pointWith l's
Length r is in one-to-one relationship；And the edge of a target is formed by connecting by a series of marginal points, therefore, withFor
Abscissa, r are ordinate, and each target can correspond to oneCurve；For an image sequence, each figure
The target of picture all respectively corresponds oneCurve, the target in whole image sequence will correspond to family of curves
Often give a specific angleFamily of curves can all correspond to one group of RiValue, this group of RiThe variance of value represents at angle
DegreeUnder identified direction, marginal point to the difference degree of centroid distance, get over by variance between the different target from image sequence
Big explanation is in angleTarget morphology variation is bigger under identified direction.Based on this, to family of curvesVariance is sought, is obtained
Variance curveVariance curve is sought laterMaximum valueCorresponding abscissa angle valueTarget morphology changes direction the most significant as in image sequence.
Detailed description of the invention
Fig. 1 is the CTPA image sequence pulmonary artery metamorphosis detection method process of the invention indicated based on radial function
Figure.
Fig. 2 is CTPA image sequence pulmonary artery target to be analyzed in the present invention, and white line indicates the edge of pulmonary artery.
Fig. 3 is target radial function representation schematic diagram in the present invention.
Fig. 4 is that the present invention indicates the family of curves to be formed to target each in Fig. 2 sequence progress radial function.
Fig. 5 is the variance curve that the present invention asks variance to obtain Fig. 4 family of curves.
Specific embodiment
Present invention will be further explained below with reference to the attached drawings and specific embodiments.
Fig. 1 is the CTPA image sequence pulmonary artery metamorphosis detection method process of the invention indicated based on radial function
Figure.
Below with reference to flow chart to the CTPA image sequence pulmonary artery metamorphosis testing principle indicated based on radial function
It is described in detail.
(1) it for the pulmonary artery sequence obtained using the target following split plot design based on edge, is indicated using radial function
Method seeks the radial function of each single pulmonary artery targetCurve.
The centroid coordinate for enabling target isIf the coordinate of object edge point is (xi,yi), i=1,2 ..., (N is N
The number of marginal point), then object edge point to centroidWire length riWith the angle of the line and trunnion axisPoint
It is not calculated as follows:
It is thus obtained a series of several rightNoteThe collection of composition is combined into Φi；
According to a series of several right?Place carries out linear interpolation, wherein(Represent nature manifold)；Obtain a series of arraysThus it drawsCurve obtains the radial function curve of single pulmonary artery target；Wherein specific step is as follows for linear interpolation:
If the coordinate value at interpolation isIn set ΦiIn find withTwo neighbouring valuesWithWhereinFor setMaximum value,For setMinimum
Value；WithCorresponding ordinate value is respectively ri1And ri2, then rj0It determines as the following formula:
(2) for entire pulmonary artery sequence image, radial function representation method in (1) is utilized to find out each pulmonary artery target
'sCurve obtains family of curvesIt takes in (1)By each givenObtain a class value R of family of curvesj, to this
Group RjVariance is sought, is obtainedBy a series of arraysObtain varianceCurve；
(3) variance for being acquired in (2)Curve acquires maximum valueThenCorresponding abscissa valueTarget morphology changes angle direction the most significant as in image sequence.
Fig. 2 is CTPA image sequence pulmonary artery target to be analyzed in the present invention, and white line indicates the edge of pulmonary artery.
Fig. 3 is target radial function representation schematic diagram in the present invention.The elliptical object as shown in figure (a), object edge point
It is with the line of centroid and the angle of trunnion axisLength is r；For star convex target, eachThere is unique r
It is corresponding；For the star convex target as shown in figure (a), the radial function curve drawn is as shown in figure (b).
Fig. 4 is that the present invention indicates the family of curves to be formed to target each in Fig. 2 sequence progress radial function.As shown in figure 4, bent
Line race density is unevenly distributed, and when curve distribution comparatively dense illustrates givenObject variations are smaller under identified direction, bent
Illustrate when line is distributed sparse givenObject variations are larger under identified direction.
Fig. 5 is the curve that the present invention asks variance to obtain Fig. 4 family of curves.The variance maximum value of the curve is 52.22, institute
Corresponding ordinate is 253 °, and the angle direction the most significant of target sequence metamorphosis shown in explanatory diagram 2 is 253 °.
Claims (1)
1. the CT image sequence pulmonary artery metamorphosis detection method indicated based on radial function, it is characterised in that: this method packet
The radial function curve using target is included to describe the form of single pulmonary artery target；It is retouched using cluster radial function family of curves
State entire pulmonary artery target sequence；Pulmonary artery series modality variation is portrayed using the variance curve of radial function race；It utilizes
Variance curve maximum value changes angle direction the most significant to screen target morphology in pulmonary artery sequence；
Wherein, the detailed step that single pulmonary artery target is described using radial function representation method is as follows:
(1) for the single pulmonary artery in pulmonary artery target sequence to be analyzed, the centroid coordinate of target is enabled to be
(2) edge of target is formed by connecting by series of points；If the coordinate of object edge point is (xi,yi), i=1,2 ...,
N, wherein N is the number of marginal point, then object edge point to centroidWire length riWith the line and trunnion axis
AngleIt is calculated as follows respectively:
Thus it is a series of several right to obtainNoteThe collection of composition is combined into Φi；
(3) a series of several right according to being obtained in (2)?Place carries out linear interpolation, wherein Represent nature manifold；Obtain a series of arraysThus it drawsCurve obtains the radial function curve of single pulmonary artery target；Wherein specific step is as follows for linear interpolation:
If the coordinate value at interpolation isIn set ΦiIn find withTwo neighbouring valuesWithIts
InFor setMaximum value,For setMinimum value；
WithCorresponding ordinate value is respectively ri1And ri2, then rj0It determines as the following formula:
Entire pulmonary artery target sequence and the variance using radial function race are described using cluster radial function family of curvesCurve portrays the detailed step of pulmonary artery series modality variation are as follows:
Each single pulmonary artery is acquired using radial function representation methodCurve obtains song for entire pulmonary artery sequence
Line raceIt takes Nature manifold is represented, by each given angle?
To a class value R of family of curvesj, to this group of RjVariance is sought, is obtainedBy a series ofObtain varianceIt is bent
Line；
Target morphology in pulmonary artery sequence is screened using variance curve maximum value changes the detailed of angle direction the most significant
Step are as follows:
For varianceCurve acquires its maximum valueThenCorresponding abscissa valueAs mesh in image sequence
Mark metamorphosis angle direction the most significant.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811601713.8A CN109727250B (en) | 2018-12-26 | 2018-12-26 | CT image sequence pulmonary artery morphological change detection method based on radial function representation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811601713.8A CN109727250B (en) | 2018-12-26 | 2018-12-26 | CT image sequence pulmonary artery morphological change detection method based on radial function representation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109727250A true CN109727250A (en) | 2019-05-07 |
| CN109727250B CN109727250B (en) | 2020-10-27 |
Family
ID=66296356
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811601713.8A Active CN109727250B (en) | 2018-12-26 | 2018-12-26 | CT image sequence pulmonary artery morphological change detection method based on radial function representation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109727250B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112465749A (en) * | 2020-11-11 | 2021-03-09 | 沈阳东软智能医疗科技研究院有限公司 | Method and device for extracting pulmonary embolism image, storage medium and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105844637A (en) * | 2016-03-23 | 2016-08-10 | 西安电子科技大学 | Method for detecting SAR image changes based on non-local CV model |
| CN107367718A (en) * | 2017-07-14 | 2017-11-21 | 河南科技大学 | A kind of multi-scatter motor-driven random hypersurface extension Target Modeling method under measuring |
| US20180204111A1 (en) * | 2013-02-28 | 2018-07-19 | Z Advanced Computing, Inc. | System and Method for Extremely Efficient Image and Pattern Recognition and Artificial Intelligence Platform |
-
2018
- 2018-12-26 CN CN201811601713.8A patent/CN109727250B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180204111A1 (en) * | 2013-02-28 | 2018-07-19 | Z Advanced Computing, Inc. | System and Method for Extremely Efficient Image and Pattern Recognition and Artificial Intelligence Platform |
| CN105844637A (en) * | 2016-03-23 | 2016-08-10 | 西安电子科技大学 | Method for detecting SAR image changes based on non-local CV model |
| CN107367718A (en) * | 2017-07-14 | 2017-11-21 | 河南科技大学 | A kind of multi-scatter motor-driven random hypersurface extension Target Modeling method under measuring |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112465749A (en) * | 2020-11-11 | 2021-03-09 | 沈阳东软智能医疗科技研究院有限公司 | Method and device for extracting pulmonary embolism image, storage medium and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109727250B (en) | 2020-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yao et al. | Label-free segmentation of COVID-19 lesions in lung CT | |
| CN107622492B (en) | Lung fissure segmentation method and system | |
| CN106780518B (en) | A kind of MR image three-dimensional interactive segmentation method of the movable contour model cut based on random walk and figure | |
| WO2018023917A1 (en) | Method and system for extracting lower limb blood vessel | |
| CN106097305B (en) | The intratracheal tree dividing method that two-pass region growing combining form is rebuild | |
| CN107977952A (en) | Medical image cutting method and device | |
| Camilus et al. | Pectoral muscle identification in mammograms | |
| CN105139377B (en) | A kind of fast robust automatic division method of abdominal CT sequence image liver | |
| US8873819B2 (en) | Method for sorting CT image slices and method for constructing 3D CT image | |
| CN101065776A (en) | Multi-component vessel segmentation | |
| CN115661149B (en) | Lung image processing system based on lung tissue data | |
| Fan et al. | Lung nodule detection based on 3D convolutional neural networks | |
| CN103324934B (en) | Blood vessel center line automatic extraction method based on parallel construction detection with cluster | |
| CN109410166A (en) | Full-automatic partition method for pulmonary parenchyma CT image | |
| Heyde et al. | Anatomical image registration using volume conservation to assess cardiac deformation from 3D ultrasound recordings | |
| CN107481215A (en) | CT Lung neoplasm dividing methods based on spatial distribution | |
| CN108364297A (en) | Blood-vessel image dividing method, terminal, storage medium | |
| Li et al. | Automatic quantification of epicardial adipose tissue volume | |
| WO2015143435A1 (en) | Graph search using non-euclidean deformed graph | |
| Metzen et al. | Matching of anatomical tree structures for registration of medical images | |
| JP2009045286A (en) | Medical image processor, medical image diagnostic apparatus, and program | |
| CN109727250A (en) | The CT image sequence pulmonary artery metamorphosis detection method indicated based on radial function | |
| Wu et al. | Automatic lung segmentation in CT images using dilated convolution based weighted fully convolutional network | |
| CN109410181A (en) | A kind of segmenting cardiac images method and device | |
| Qiao et al. | Automatic liver segmentation method based on improved region growing algorithm |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |