CN109754400B - Vein removal method - Google Patents
Vein removal method Download PDFInfo
- Publication number
- CN109754400B CN109754400B CN201910051735.XA CN201910051735A CN109754400B CN 109754400 B CN109754400 B CN 109754400B CN 201910051735 A CN201910051735 A CN 201910051735A CN 109754400 B CN109754400 B CN 109754400B
- Authority
- CN
- China
- Prior art keywords
- vein
- width
- coronary artery
- length
- candidate
- 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
Images
Abstract
The invention discloses a vein removal method, which comprises the following steps: s1, obtaining an initial 3D coronary artery segmentation body; s2, setting the length and the step length of a sliding window, and starting to detect the width abnormity of the initial 3D coronary artery segmentation body from the starting point of the central line of the initial 3D coronary artery segmentation body; s3, intercepting the position interval with abnormal width, and performing confidence adjustment and segmentation to obtain a maximum segmentation body; s4, replacing the coronary artery with the maximum segmentation body in the abnormal interval with the corresponding width position; s5, performing bifurcation point detection on the 3D coronary artery segmentation body which completes width abnormality detection; and S6, judging the abnormality of the branch point, and identifying and removing the vein. The invention can identify and remove two kinds of common vein noises (bonding and crossing), and further output the coronary artery segmentation result after vein interference is eliminated, and the result is used as the basis of the next coronary artery reconstruction so as to improve the precision of the coronary artery reconstruction.
Description
Technical Field
The invention relates to the field of coronary image processing, in particular to a vein removing method based on morphological recognition.
Background
The automatic coronary reconstruction has important clinical value and practical significance for doctors. The coronary artery segmentation is an important ring for automatic coronary artery reconstruction, and the accuracy of the segmentation result directly influences the accuracy of the result of subsequent automatic coronary artery reconstruction.
In the coronary segmentation process, veins are usually the main noise, and the veins and arteries are mainly distributed in two forms: 1. the two are attached or adhered in parallel; 2. the two are crossed. The characteristic forms of veins are difficult to distinguish from arteries, so that the veins are difficult to remove.
Disclosure of Invention
The invention aims to provide a method for identifying and removing veins with two morphologically distributed modes respectively.
In order to achieve the purpose, the invention adopts the following technical scheme:
a vein removal method, comprising:
s1, obtaining an initial 3D coronary artery segmentation body;
s2, setting the length and the step length of a sliding window, and carrying out width abnormity detection on the initial 3D coronary artery segmentation body from the starting point of the central line of the initial 3D coronary artery segmentation body:
s21, determining the reference width: sliding a plurality of windows forwards by taking the current window as a reference, and determining the average value of the coronary artery widths of all the sliding windows as a reference width if the coronary artery width variation value is smaller than a set value;
s22, calculating the ratio of the coronary artery width of the current window to the reference width, and marking the ratio as a width abnormal position when the ratio is larger than a set threshold;
s3, intercepting the position interval with abnormal width, adjusting and dividing confidence coefficient to obtain the maximum division body:
s31, sliding the window, and taking out the head and tail positions of the width abnormal position to obtain a width abnormal position interval;
s32, increasing confidence coefficient of segmentation output corresponding to the width abnormal position interval until the reference width of the width abnormal position interval appears;
s33, keeping the maximum division body, and discarding the rest division bodies;
and S4, replacing the corresponding coronary artery with the abnormal interval with the maximum segmentation body.
Further, the number of sliding windows of the reference width is determined to be 5.
Further, the window length is 10, and the step size is half of the window length.
Further, the threshold value of the width abnormality position is determined to be 1.6.
Further, the method also comprises the following steps:
s5, performing bifurcation point detection on the 3D coronary artery segmentation body which completes width abnormality detection;
s6, carrying out abnormity judgment on the branch point, and identifying and removing veins:
s61, according to the center point of the preamble, identifying the artery and the branching venation, and marking the branching venation as a candidate vein;
s62, identifying candidate vein numbers of each bifurcation point to adopt different strategies for identification:
when the number of candidate veins is 1: calculating an included angle between the approximate trend of the artery and the approximate trend of the candidate vein, and identifying the vein when the included angle is larger than alpha; α +90 ° β, β being the tolerance angle;
when the number of candidate veins is 2: judging a class straight line of each candidate vein, if the class straight line is the class straight line, identifying the candidate vein as the vein, and otherwise, identifying the candidate vein by referring to an identification method when the number of the candidate veins is 1;
and S63, removing the forked venation identified as the vein, and outputting the corrected 3D coronary artery segmentation body.
Further, β is 15 °.
Further, the approximate trend of the artery is calculated by:
firstly, selecting an artery with the same length as the candidate vein by taking a bifurcation point as an end point;
secondly, setting the maximum length and the minimum length of the available central line of the artery to form an available central line length interval;
and finally, searching curve segments with the change rate smaller than a set threshold value in the length interval of the available central line, and calculating the average vector direction of the curve segments as the approximate trend of the artery.
Further, the maximum length is 60 pixel points, and the minimum length is 25 pixel points.
Further, the set threshold is 0.15.
After adopting the technical scheme, compared with the background technology, the invention has the following advantages: the invention can identify and remove two kinds of common vein noises (bonding and crossing), and further outputs the coronary artery segmentation result after vein interference is eliminated, and the result is used as the basis of the next coronary artery reconstruction.
Drawings
FIG. 1 is a schematic flow diagram of the present invention;
FIG. 2 is a schematic diagram of width anomaly detection according to the present invention;
FIG. 3 is a schematic diagram of confidence adjusted segmentation of the present invention, where FIG. 3(a) is a schematic diagram with a confidence of 0.5 and FIG. 3(b) is a schematic diagram with a confidence of 0.9;
fig. 4 is a schematic diagram of coronary bifurcation of the present invention, fig. 4(a) is a schematic diagram of 3-bifurcation, and fig. 4(b) is a schematic diagram of 4-bifurcation.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Examples
Referring to fig. 1, the present invention discloses a vein removal method, which specifically includes:
and S1, carrying out segmentation processing on the coronary image to obtain an initial 3D coronary segmentation body.
S2, setting the length and the step length of a sliding window, and carrying out width abnormity detection on the initial 3D coronary artery segmentation body from the starting point of the central line of the initial 3D coronary artery segmentation body:
s21, determining the reference width: sliding a plurality of windows forwards by taking the current window as a reference, and determining the average value of the coronary artery widths of all the sliding windows as a reference width if the coronary artery width variation value is smaller than a set value;
s22, calculating the ratio of the coronary artery width of the current window to the reference width, and marking the position with abnormal width when the ratio is larger than the set threshold value
In this embodiment, the number of consecutive windows for determining the reference width is 5. The window length is 10, the step length is half of the window length, and the threshold value for determining the width abnormal position is 1.6.
If a width anomaly is encountered during the sliding window (width discontinuity greater than 1.6), it can be handled in two ways:
1. the reference width value determined by the previous window is the reference width value of the current window;
2. and forward calculation, namely ignoring the window with abnormal width, and taking the width of the window at the position as the basis for calculating the reference width when the similar width before the abnormal occurs after the abnormal position.
As shown in fig. 2, the direction of the arrow is the detection direction, and it can be seen that initially, the width of the coronary artery in the frame is approximately the same, and thereafter, the width of the coronary artery in the frame sequentially undergoes widening, holding, narrowing until it is the same as or narrower than the width of the coronary artery before the change.
S3, intercepting the position interval with abnormal width, adjusting and dividing confidence coefficient to obtain the maximum division body:
and S31, sliding the window, and taking out the head and tail positions of the width abnormal position to obtain a width abnormal position interval.
And S32, increasing confidence coefficient of segmentation output corresponding to the position interval with abnormal width until the position interval with abnormal width has reference width, and then considering that the noise is separated from the blood vessel.
As shown in fig. 3, the gray part of the graph represents a blood vessel, the black line represents a center line, the confidence is 0.5 when the graph is in a stuck state as shown in (a), and the confidence is 0.9 when the graph is in a separated state as shown in (b).
S33, keeping the largest division body, and discarding the rest division bodies as interference blocks.
And S4, replacing the corresponding coronary artery with the abnormal interval with the maximum segmentation body.
In this way, the vein noise treatment of the bonding site can be completed, but the vein noise has another embodiment form, namely, the vein noise intersects with the artery, so that the vein noise needs to be identified and removed.
And S5, detecting bifurcation points of the initial 3D coronary artery segmentation body.
And S6, judging the abnormality of the branch point, and identifying and removing the vein.
And S61, identifying the artery and the branching venation according to the preorder central point in the 3D coronary artery segmentation body.
As shown in fig. 4(a) and 4(b), the computer identifies the preorder center point in the initial 3D coronary segment, and runs along preorder center points a1, a2, a3, the venation of which is an artery, and the branching venation extending from the center point a1 is the branching venation, and marks the branching venation as a candidate vein.
S62, where the normal artery has a cross, the normal artery is divided into two categories, one is 3-way and the other is 4-way, so that the candidate venous number of each bifurcation point needs to be identified to adopt different strategies to identify:
as shown in fig. 4(a), when the candidate vein number is 1 (i.e., 3-way state): and calculating the included angle between the approximate trend of the artery and the approximate trend of the candidate vein (namely the fitting straight line of the central line of the segment), and identifying the vein and marking when the included angle is larger than alpha.
The coronary artery of a person is basically a tree-shaped downward branch, the branch generally does not appear to be long upwards, namely, the situation that alpha is larger than 90 degrees does not normally appear, and when the alpha is larger than 90 degrees, the possibility that the branch is a vein is high, so that the branch can be judged as the vein.
In consideration of individual differences, the present embodiment is verified by a large number of samples, and a tolerance angle β is set on the basis of 90 °, β is preferably 15 °, and α is 90 ° +15 ° + 105 °.
When the number of candidate veins is 2 (i.e., 4-way state), as shown in fig. 4 (b): and judging the class straight line of each candidate vein, identifying the candidate vein as a vein and marking if the candidate vein is the class straight line, and otherwise, identifying by referring to an identification method when the number of the candidate veins is 1.
And judging the similar straight line, namely fitting the central line point set of the section, and judging the similar straight line if better fitting property is obtained after fitting. It can be achieved by at least two constraints: 1. the fitting error is small, and when the fitting error value is smaller than a set value, the straight line can be judged to be a similar straight line; 2. and the distance between any central line point and the fitting line is less than a set threshold value, and the line is judged to be a straight-like line. The second constraint is adopted in the present embodiment, and the set threshold is 0.15.
The approximate trend of the artery is calculated by:
firstly, calculating the length of a candidate vein, and selecting an artery with the same length as the candidate vein by taking a bifurcation point as an end point;
secondly, setting the maximum length and the minimum length of the available central line of the artery to form an available central line length interval, wherein the available central line length interval is [25,60] in the embodiment;
and finally, searching curve segments with the change rate smaller than a set threshold value in the length interval of the available central line, and calculating the average vector direction of the curve segments as the approximate trend of the artery.
The rate of change is calculated by: d is the distance from any point on the target line segment to the head-tail connecting line of the target line segment, and L is the length of the head-tail connecting line of the target line segment. In this embodiment, the set threshold is 0.15.
And S63, removing the forked venation identified as the vein, and outputting the modified 3D coronary artery segmentation body.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (9)
1. A vein removal method, comprising:
s1, obtaining an initial 3D coronary artery segmentation body;
s2, setting the length and the step length of a sliding window, and carrying out width abnormity detection on the initial 3D coronary artery segmentation body from the starting point of the central line of the initial 3D coronary artery segmentation body:
s21, determining the reference width: sliding a plurality of windows forwards by taking the current window as a reference, and determining the average value of the coronary artery widths of all the sliding windows as a reference width if the coronary artery width variation value is smaller than a set value;
s22, calculating the ratio of the coronary artery width of the current window to the reference width, and marking the ratio as a width abnormal position when the ratio is larger than a set threshold;
s3, intercepting the position interval with abnormal width, adjusting and dividing confidence coefficient to obtain the maximum division body:
s31, sliding the window, and taking out the head and tail positions of the width abnormal position to obtain a width abnormal position interval;
s32, increasing confidence coefficient of segmentation output corresponding to the width abnormal position interval until the reference width of the width abnormal position interval appears;
s33, keeping the maximum division body, and discarding the rest division bodies;
and S4, replacing the corresponding coronary artery with the abnormal interval with the maximum segmentation body.
2. The vein removal method of claim 1, wherein: the number of sliding windows for the reference width is determined to be 5.
3. The vein removal method of claim 1, wherein: the window length is 10 and the step length is half of the window length.
4. The vein removal method of claim 1, wherein: the threshold value for the width anomaly location was determined to be 1.6.
5. The vein removal method of any one of claims 1-4, further comprising the subsequent steps of:
s5, performing bifurcation point detection on the 3D coronary artery segmentation body which completes width abnormality detection;
s6, carrying out abnormity judgment on the branch point, and identifying and removing veins:
s61, according to the center point of the preamble, identifying the artery and the branching venation, and marking the branching venation as a candidate vein;
s62, identifying candidate vein numbers of each bifurcation point to adopt different strategies for identification:
when the number of candidate veins is 1: calculating an included angle between the approximate trend of the artery and the approximate trend of the candidate vein, and identifying the vein when the included angle is larger than alpha; α = β +90 °, β is the tolerated angle;
when the number of candidate veins is 2: judging a class straight line of each candidate vein, if the class straight line is the class straight line, identifying the candidate vein as the vein, and otherwise, identifying the candidate vein by referring to an identification method when the number of the candidate veins is 1;
and S63, removing the forked venation identified as the vein, and outputting the corrected 3D coronary artery segmentation body.
6. The vein removal method of claim 5, wherein: β =15 °.
7. The vein removal method of claim 5, wherein: the approximate trend of the artery is calculated by:
firstly, selecting an artery with the same length as the candidate vein by taking a bifurcation point as an end point;
secondly, setting the maximum length and the minimum length of the available central line of the artery to form an available central line length interval;
and finally, searching curve segments with the change rate smaller than a second set threshold value in the length interval of the available central line, and calculating the average vector direction of the curve segments as the approximate trend of the artery.
8. The vein removal method of claim 7, wherein: the maximum length is 60 pixel points, and the minimum length is 25 pixel points.
9. The vein removal method of claim 7, wherein: the second set threshold is 0.15.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910051735.XA CN109754400B (en) | 2019-01-21 | 2019-01-21 | Vein removal method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910051735.XA CN109754400B (en) | 2019-01-21 | 2019-01-21 | Vein removal method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109754400A CN109754400A (en) | 2019-05-14 |
| CN109754400B true CN109754400B (en) | 2020-12-29 |
Family
ID=66405867
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910051735.XA Active CN109754400B (en) | 2019-01-21 | 2019-01-21 | Vein removal method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109754400B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103099634A (en) * | 2011-10-14 | 2013-05-15 | 美国西门子医疗解决公司 | Method and apparatus for correcting medical imaging data |
| CN107248155A (en) * | 2017-06-08 | 2017-10-13 | 东北大学 | A kind of Cerebral venous dividing method based on SWI images |
| CN108492309A (en) * | 2018-01-21 | 2018-09-04 | 西安电子科技大学 | Magnetic resonance image medium sized vein blood vessel segmentation method based on migration convolutional neural networks |
| CN108932714A (en) * | 2018-07-23 | 2018-12-04 | 苏州润心医疗器械有限公司 | The patch classification method of coronary artery CT image |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9848785B2 (en) * | 2013-12-05 | 2017-12-26 | Siemens Healthcare Gmbh | Analysis and characterization of patient signals |
-
2019
- 2019-01-21 CN CN201910051735.XA patent/CN109754400B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103099634A (en) * | 2011-10-14 | 2013-05-15 | 美国西门子医疗解决公司 | Method and apparatus for correcting medical imaging data |
| CN107248155A (en) * | 2017-06-08 | 2017-10-13 | 东北大学 | A kind of Cerebral venous dividing method based on SWI images |
| CN108492309A (en) * | 2018-01-21 | 2018-09-04 | 西安电子科技大学 | Magnetic resonance image medium sized vein blood vessel segmentation method based on migration convolutional neural networks |
| CN108932714A (en) * | 2018-07-23 | 2018-12-04 | 苏州润心医疗器械有限公司 | The patch classification method of coronary artery CT image |
Non-Patent Citations (2)
| Title |
|---|
| 基于Hessian矩阵的手指静脉图像分割;林剑 等;《机电工程》;20161231;第33卷(第12期);第1517-1522页 * |
| 眼底图像动静脉宽度差变化率测量方法;郭莹 等;《仪器仪表学报》;20160430;第37卷(第4期);第912-919页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109754400A (en) | 2019-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chutatape et al. | Retinal blood vessel detection and tracking by matched Gaussian and Kalman filters | |
| CN110599539B (en) | Stripe center extraction method of structured light stripe image | |
| CN111681224A (en) | Method and device for acquiring blood vessel parameters | |
| CN107169979A (en) | A kind of method for detecting image edge of improvement Canny operators | |
| JP2004508623A (en) | Image processing method and system for extracting continuous points following a linear structure in an image sequence | |
| CN110659645B (en) | Character recognition method for digital instrument | |
| CN110197133B (en) | Method and apparatus for aligning detected waveforms | |
| CN113349737B (en) | Calibration method of OCT (optical coherence tomography) image of intravascular dual-mode imaging system | |
| CN109754400B (en) | Vein removal method | |
| CN107369164B (en) | Infrared weak and small target tracking method | |
| JP2019532789A5 (en) | ||
| JP3821739B2 (en) | Measurement data shaping method | |
| CN108898148A (en) | A kind of digital picture angular-point detection method, system and computer readable storage medium | |
| CN109800711B (en) | Vein removal method based on morphological recognition | |
| CN117454096B (en) | Motor production quality detection method and system | |
| CN116616804B (en) | Method, device, equipment and storage medium for acquiring intracranial arterial stenosis evaluation parameters | |
| CN116823820B (en) | Aluminum-carbon integral stopper rod contour bending detection method | |
| CN109875595B (en) | Intracranial vascular state detection method and device | |
| CN110236594B (en) | Fetal movement signal identification marking method, system, medium and equipment | |
| CN111325789B (en) | Curvature discontinuous point detection method based on discrete direction change sequence | |
| CN105787444B (en) | Signal denoising method based on V system | |
| CN110021016B (en) | Calcification detection method | |
| CN110033466B (en) | Coronary artery straightening image segmentation boundary determination method based on multiple gray levels | |
| CN110288569B (en) | Crack near-field deformation analysis area determination method based on level set method | |
| Ghanaei et al. | Automatic graph-based method for classification of retinal vascular bifurcations and crossovers |
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 | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 100120 rooms 303, 304, 305, 321 and 322, building 3, No. 11, Chuangxin Road, science and Technology Park, Changping District, Beijing Patentee after: Shukun (Beijing) Network Technology Co.,Ltd. Address before: Unit 547, unit 1, building 1, yard 1, Longyu middle street, Huilongguan town, Changping District, Beijing Patentee before: SHUKUN (BEIJING) NETWORK TECHNOLOGY Co.,Ltd. |