CN108898626B - A kind of autoegistration method coronarius - Google Patents
A kind of autoegistration method coronarius Download PDFInfo
- Publication number
- CN108898626B CN108898626B CN201810643770.6A CN201810643770A CN108898626B CN 108898626 B CN108898626 B CN 108898626B CN 201810643770 A CN201810643770 A CN 201810643770A CN 108898626 B CN108898626 B CN 108898626B
- Authority
- CN
- China
- Prior art keywords
- point
- coronary artery
- bifurcation
- centerline
- image
- 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
- 238000000034 method Methods 0.000 title claims abstract description 55
- 210000004351 coronary vessel Anatomy 0.000 claims abstract description 126
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 18
- 230000011218 segmentation Effects 0.000 claims abstract description 14
- 239000013598 vector Substances 0.000 claims description 31
- 239000004576 sand Substances 0.000 claims description 25
- 230000009466 transformation Effects 0.000 claims description 25
- 238000005070 sampling Methods 0.000 claims description 22
- 239000011159 matrix material Substances 0.000 claims description 11
- 238000010968 computed tomography angiography Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000003709 image segmentation Methods 0.000 claims description 3
- 101100421536 Danio rerio sim1a gene Proteins 0.000 claims 1
- 101100495431 Schizosaccharomyces pombe (strain 972 / ATCC 24843) cnp1 gene Proteins 0.000 claims 1
- 239000008280 blood Substances 0.000 claims 1
- 210000004369 blood Anatomy 0.000 claims 1
- 239000000284 extract Substances 0.000 abstract description 4
- 238000012545 processing Methods 0.000 abstract description 2
- 238000012217 deletion Methods 0.000 abstract 1
- 230000037430 deletion Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 9
- 210000002216 heart Anatomy 0.000 description 9
- 238000003745 diagnosis Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 4
- 238000002591 computed tomography Methods 0.000 description 3
- 230000000747 cardiac effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000012634 fragment Substances 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 208000032594 Vascular Remodeling Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000002583 angiography Methods 0.000 description 1
- 210000000709 aorta Anatomy 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 208000029078 coronary artery disease Diseases 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 231100000518 lethal Toxicity 0.000 description 1
- 230000001665 lethal effect Effects 0.000 description 1
- 210000004072 lung Anatomy 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000000241 respiratory effect Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- 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
-
- 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
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The present invention proposes a kind of autoegistration method coronarius, is related to field of medical image processing.This method obtains a pair of of coronary artery images of acquisition first, carries out blood vessel segmentation to every image and extracts coronary artery center line.Then it to the central line pick-up bifurcated point feature in every image and matches, obtains final matched bifurcation point to set；Using the set, center line segment is registrated, peculiar bifurcation and segment centerline segment are deleted, the sampled point after center line segment pair and fine registration after being matched is to set；To the peculiar center line segment of deletion, corresponding omission segment is further divided and is registrated in another image, obtains the registration result of final coronary artery center line.The present invention can handle two inconsistent situations of coronary artery center line topological structure, and be further registrated to omission segment coronarius and center line segment, and the integrality of segmentation and registration result is improved.
Description
Technical Field
The invention relates to the field of medical image processing, in particular to an automatic registration method of coronary arteries.
Background
Coronary artery disease is one of the most lethal factors worldwide, and computed tomography angiography is the mainstream method of coronary artery imaging. By comparing images of the coronary arteries of the same patient at different times (e.g., initial visit and review), the progression of the disease can be observed to facilitate adjustments to the treatment plan. By comparing the coronary artery images of different phases of the heart movement cycle, the movement rule and the lumen change rule of the coronary artery in the heart cycle can be analyzed. There are differences in shape, posture, etc. between different coronary images, which are caused both by external causes (e.g., different points in time during the heart motion cycle for image acquisition, different relative positions and angles of the body and the instrument during imaging) and internal causes (e.g., possible vascular remodeling, patient heart and respiratory motion). The difference causes difficulty in contrast analysis of the coronary artery, so that accurate automatic registration is performed on the coronary artery in advance, the corresponding relation of each point is determined, and the necessity is high, so that the diagnosis efficiency and the diagnosis accuracy can be greatly improved.
Coronary arteries have a tree-like structure with many branches and are distributed almost on the surface of the whole heart, and the lumens are fine. In view of its tree-like structure, coronary vessels are generally viewed as a collection of vessel segments, a vessel segment being defined as the portion of a vessel sandwiched between two bifurcation points or a bifurcation point and an end point (beginning or end). Most of the existing medical image registration methods are directed to organs with large volumes and concentrated distribution, such as brain, heart and lung. If a general image registration method is applied to the coronary artery, the registration effect is poor due to the severe influence of other surrounding tissues. It is necessary to design a registration method for coronary arteries with high accuracy.
The existing registration method for coronary artery includes extracting the central lines of coronary artery blood vessels from a pair of computed tomography angiography images (where a pair of images refers to two images obtained by two scans of the same patient at a time interval or two images at different time points of the cardiac cycle in the same scan), and performing intensive point sampling on the two central lines to obtain two sets of point sets, and using the two sets of point sets as the input of the registration method. Then, the two point sets are matched by using a coherent point drift method, and a deformation field obtained by using the point set matching calculation is used as a space transformation model for coronary artery registration. Such a method has disadvantages including: 1) the discrete points on the centerline of the coronary artery of the whole tree structure are matched as an integral point set, the points on different branches are not distinguished, and the mismatching that the points on two different branches are drawn close is easy to occur; 2) two tree-like coronary artery centerlines as inputs tend to have topology inconsistency situations, such as where one coronary artery has branches that are missing in the other coronary artery, the method takes all points on the coronary artery as a whole, so point matching and possible errors where the topology is inconsistent occur; 3) the registration of this method only focuses on the common part of the two coronary arteries, which is chosen to be directly ignored, even if no erroneous point matching can occur, for vessels that one coronary artery has and the other one is missing. However, these vessels are likely to have lesions and are particularly important for diagnostic and medical research, and direct omission greatly reduces the value of practical application.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides an automatic registration method of coronary arteries. The invention can process the situation that the topological structures of the central lines of the two input coronary arteries are not consistent, can further segment and extract the missed vessel segments and the central line segments of the input coronary arteries, can perform further registration, improves the integrity of the segmentation and registration results, and has more practical medical research and diagnosis values.
A method for automatic registration of coronary arteries, characterized in that it comprises the following steps:
1) acquiring coronary artery images, performing blood vessel segmentation on each coronary artery image and extracting a corresponding central line; the method comprises the following specific steps:
1.1) acquiring a coronary artery image;
acquiring a pair of computed tomography angiography images which are performed aiming at a certain coronary artery, and randomly selecting one of the two images as a source image and the other image as a target image;
1.2) segmenting coronary artery blood vessels for each image obtained in the step 1.1), and extracting a central line corresponding to each coronary artery blood vessel; the coronary artery blood vessel in the source image is recorded as VsThe center line is marked as Cs(ii) a Coronary artery vessels in the target image are registered as VtThe center line is marked as Ct；
2) Extracting bifurcation point characteristics from the central line in each coronary artery image in the step 1) and matching to obtain a final matched bifurcation point pair set; the method comprises the following specific steps:
2.1) extracting the features of the bifurcation points;
let BsRepresenting the whole n of the coronary artery central lines in the source imagesA point set of points of bifurcation, BtRepresenting the whole of the coronary centerline n in the target imagetA point set consisting of multiple bifurcation points; extracting a corresponding feature descriptor for a bifurcation point of each coronary artery centerline, and representing the descriptor in a vector form; wherein the descriptor vector of the ith bifurcation point of the source image is taken asThe descriptor vector of the ith bifurcation point in the target image is taken as
2.2) obtaining a candidate matching bifurcation point pair set Acand；
Calculation of BtEach branch point b intAnd BsThe Euclidean distance of the descriptor vectors of all the bifurcation points is calculated, and B corresponding to the minimum Euclidean distance in the result is calculatedsThe bifurcation point in (1) is denoted by btCorresponding bifurcation point bsIn total, n is obtainedtA point pair of bifurcations, ntThe set of the branch point pairs is marked as a candidate matching branch point pair set Acand；
2.3) computing the set AcandThe attribute matrix M of (2);
first, wait AcandA graph G (V, E, M) is established, wherein each node in the set of nodes V of the graph corresponds to AcandEach side in the side set E of the graph corresponds to the relationship between every two bifurcation points, the off-diagonal element M (i, j) of the attribute matrix M reflects the compatibility between the ith bifurcation point pair and the jth bifurcation point pair, and the diagonal element M (i, i) of the attribute matrix M represents the descriptor sub-vector similarity of the two bifurcation points in the ith bifurcation point pair; the expression of M is as follows:
wherein d isiRepresenting two bifurcation points m in the ith bifurcation point pairi1And mi2Euclidean distance of thetaiIs mi1And mi2The relative angle of the directions is such that,is a handle mi2M when the corresponding vector is taken as the polar axisi1The polar angle of (d); THdist、THθAndthreshold values, | θ, representing lower limits of ratio (i, j), respectivelyi-θjThreshold sum of | upper boundA threshold value of an upper limit; THdistTake on the value of [0.4, 0.9]Interval, THθTake on [30 degrees, 90 degrees ]]The interval of time is,take on [30 degrees, 90 degrees ]]An interval;
2.4) calculating the final matched bifurcation point pair set Afinal(ii) a The method comprises the following specific steps:
2.4.1) establishing a final matched bifurcation point pair set AfinalAnd A isfinalInitializing to an empty set;
2.4.2) decomposing the eigenvalue of the matrix M to find the eigenvector x corresponding to the maximum eigenvalue of M*；
2.4.3) for x*The elements in the sequence are sorted from big to small to obtain the sequence x ', and the elements of the sequence x' from front to back are arranged in the sequence x*The sequence numbers in (1) form a sequence L;
2.4.4) determine L: if L is null, then output the current AfinalAnd ending the solution; if L is not empty, entering step 2.4.5);
2.4.5) takes the first value L (1) in the sequence L and decides: if x*(L (1)) < ε, the current A is outputfinalAnd ending the solution, wherein the value of epsilon is [0.0000001, 0.01 ]]An interval; otherwise, entering step 2.4.6);
2.4.6) if AcandThe L (1) th branch point pair in (A)finalIf any pair of the branch points contains the same branch point, deleting L (1) from L, and returning to the step 2.4.4); otherwise, go to step 2.4.7);
2.4.7) A is treatedcandThe L (1) th branch point pair in (A) is added into the set AfinalDeleting L (1) from L, and returning to the step 2.4.4);
3) updating the central line in each coronary artery image to obtain a deleted central line segment set; matching the centerline segments by using the updated centerline and the finally matched bifurcation point pair set in the step 2); the method comprises the following specific steps:
3.1) updating the central line in each coronary artery image to obtain a deleted central line segment set; the method comprises the following specific steps:
3.1.1) obtaining CsOr CtRespectively calculating the direction vectors of two branch centerline segments and a trunk centerline segment corresponding to the special bifurcation point at the bifurcation point, deleting the branch centerline segment with a larger included angle between the direction vectors and the direction vectors of the trunk centerline segments, and deleting the special bifurcation point;
3.1.2) from CsAnd CtAnd continuously searching for a unique bifurcation point from the remaining bifurcation points and judging: if the unique bifurcation point exists, returning to the step 3.1.1) again; if no specific bifurcation point exists, ending the centerline updating process to obtain the coronary artery centerline of the updated source image as C'sAnd the coronary artery center line of the updated target image is recorded as C'tAnd recording the Set of all deleted centerline segments as Setseg_del；
3.2) matching of centerline segments;
from C 'obtained in step 3.1)'sAnd C'tAnd the final matched bifurcation point pair set A obtained in the step 2)finalCenterline segments that satisfy one of any two of the following conditions are considered a match: a) two centerline segments sandwiched between two pairs of matched bifurcation point pairs; b) one end is the starting point or the terminal point of the coronary artery, the other end is a matched bifurcation point pair, and the direction included angle of two bifurcation points of the bifurcation point pair is smaller thanTwo centerline segments of a fixed angle threshold; the Set of matched centerline segment pairs is denoted as Setseg_pair；
3.3) fine registration of centerline segments;
will Setseg_pairThe segment on the center line belonging to the source image is marked as Segs，Setseg_pairThe center line segment belonging to the target image is marked as Segt(ii) a For SegsAnd SegtPoint sampling, Seg, is performed separatelysIs a sampling interval ofs，SegtIs a sampling interval oft；
Let SegsAnd SegtRespectively have NsAnd NtSampling points, defining a discrete transformation model between a set of points as follows:
S：＝{S(i)＝j，i∈{0，1，...，Nt}，j∈{0，1，...，Ns}}，
wherein S (i) is SegtS (i) ═ j denotes the state of the ith sample point of (g)tThe ith sampling point of (1) and the SegsCorresponds to the jth sample point of (a);
the objective function is:
the invention has the characteristics and beneficial effects that:
the invention registers coronary artery in different computer tomography angiography images, firstly finds out the bifurcation point of the centerline of the coronary artery, decomposes the centerline of the tree-shaped coronary artery into segments by utilizing the bifurcation point information, finds out the matching relation between the centerline segments of the two coronary arteries, and simultaneously determines the part with inconsistent topological structure. The invention carries out point set matching on the sampling points on the paired centerline segments, so that the situation of point mismatching on different branches can not occur, and the matching process can not be influenced by inconsistent topological structures. In addition, for the vessel segments and the center line segments which are missed by one coronary artery and the other coronary artery, the invention can further segment and extract the vessel segments and perform further registration, so that a plurality of lesion vessel segments with medical value are segmented and registered, and the invention has important medical research and diagnosis values.
1) In case of the topology inconsistency of the two coronary arteries input, the registration effect of the invention is not influenced by the fact
Influence and strong robustness;
2) the invention can perform segmentation and extraction on the missed blood vessels and central line segments in the coronary artery again, and simultaneously perform segmentation and extraction on the missed blood vessels and central line segments in the coronary artery
The integrity of segmentation and registration is improved, and the part of blood vessel has strong medical research and diagnosis values;
3) the invention can reach the highest registration precision at present.
Drawings
FIG. 1 is an overall flow chart of the method of the present invention.
Fig. 2 is a schematic diagram of a bifurcation point matching related parameter according to an embodiment of the present invention.
FIG. 3 is a schematic diagram of centerlines of topological inconsistencies, according to an embodiment of the present invention.
Fig. 4 is a centerline example diagram of topology inconsistency according to an embodiment of the present invention.
Fig. 5 is a diagram of the centerline segment registration result according to an embodiment of the present invention.
FIG. 6 is a schematic diagram of a further segmented coronary vessel model according to an embodiment of the present invention.
Fig. 7 is a diagram of the coronary artery registration result according to the embodiment of the present invention.
Fig. 8 is a diagram of the coronary artery registration result according to the embodiment of the present invention.
Detailed Description
The present invention provides an automatic registration method for coronary artery, which is further described in detail below with reference to the accompanying drawings and specific embodiments.
The invention provides an automatic registration method of coronary arteries, the overall flow is shown as figure 1, and the method comprises the following steps:
1) acquiring coronary artery images, performing blood vessel segmentation on each coronary artery image and extracting a corresponding central line;
two computed tomography angiography images (coronary artery images for short) of the coronary artery are obtained, the coronary artery blood vessel is segmented from the two images respectively, and the corresponding coronary artery central line is extracted. The method comprises the following specific steps:
1.1) acquiring a coronary artery image;
a pair of computed tomography angiographic images is acquired of a coronary artery. Here, a pair of images may be images obtained from two scans of the same patient at intervals, or may be two images of the same scan at different points in time of the cardiac cycle. One of the two images is selected as a source image, and the other image is used as a target image.
1.2) segmenting coronary artery blood vessels for each image obtained in the step 1.1), and extracting a central line corresponding to each coronary artery blood vessel;
the results of vessel segmentation and centerline extraction of coronary arteries may be manually labeled in a computed tomography angiographic image, or segmented and extracted using a semi-automatic or fully-automatic algorithmIn (1). The coronary artery blood vessel in the source image is recorded as VsThe center line is marked as Cs(ii) a Coronary artery vessels in the target image are registered as VtThe center line is marked as Ct。
2) Extracting bifurcation point characteristics from the central line in each coronary artery image in the step 1) and matching to obtain a final matched bifurcation point pair set; the method comprises the following specific steps:
2.1) extracting the features of the bifurcation points;
the central line (called as "coronary artery central line" for short) corresponding to each coronary artery vessel obtained in step 1) is of a tree structure and comprises a group of bifurcation points, wherein the bifurcation points refer to points where bifurcation is generated by the coronary artery central line. Let BsRepresenting the whole n of the coronary artery central lines in the source imagesA point set of points of bifurcation, BtRepresenting the whole of the coronary centerline n in the target imagetA point set of bifurcation points.
In the present invention, the bifurcation point is endowed with both positional and directional properties. The position is represented by three-dimensional spatial coordinates, and the direction is defined as the tangential direction along the centerline of the vessel and pointing from the proximal end to the distal end of the heart, represented by a direction vector. The invention extracts a corresponding feature descriptor (such as a three-dimensional scale invariant feature transformation descriptor or other type of feature descriptor) for the bifurcation point of each coronary artery centerline, and represents the descriptor in a vector form. Wherein the descriptor vector of the ith bifurcation point of the source image is taken asThe descriptor vector of the ith bifurcation point in the target image is taken as
2.2) obtaining a candidate matching bifurcation point pair set Acand；
Calculation of BtEach of which is divided intoCross point btAnd BsThe Euclidean distance of the descriptor vectors of all the bifurcation points is calculated, and B corresponding to the minimum Euclidean distance in the result is calculatedsThe bifurcation point in (1) is denoted by btCorresponding bifurcation point bsIn total, n is obtainedtA point pair of bifurcations, ntThe set of the branch point pairs is marked as a candidate matching branch point pair set Acand。AcandThe candidate matching bifurcation point pair set is obtained preliminarily in the bifurcation point matching process, and then the final matching bifurcation point pair is selected through the steps 2.3) and 2.4).
2.3) computing the set AcandThe attribute matrix M of (2);
the final matching result of the bifurcation point can be obtained by a plurality of methods, and a point set matching method based on a spectral clustering method is taken as an example.
First, for AcandA graph G (V, E, M) is established, wherein each node in the set of nodes V of the graph corresponds to AcandEach side in the side set E of the graph corresponds to a relationship between every two bifurcation point pairs, a non-diagonal element M (i, j) (i ≠ j) of the attribute matrix M reflects compatibility between the ith bifurcation point pair and the jth bifurcation point pair (the value is in a (0, 1) interval, the larger the value is, the higher the compatibility is, and a diagonal element M (i, i) of M represents the similarity of descriptor sub-vectors of two bifurcation points in the ith bifurcation point pair (the value is in a (0, 1) interval, and the larger the value is, the higher the similarity is). M is related to three geometric parameters, as shown in FIG. 2, diRepresenting the bifurcation point m of the ith bifurcation point pairi1And mi2Euclidean distance of thetaiIs mi1And mi2The relative angle of the directions is such that,is a handle mi2M when the corresponding vector is taken as the polar axisi1The polar angle of (1). The specific definition of M is as follows:
wherein TH isaist、THθAndthreshold values, | θ, representing lower limits of ratio (i, j), respectivelyi-θjThreshold sum of | upper bound The upper threshold. THdistTake on the value of [0.4, 0.9]Interval, THθTake on [30 degrees, 90 degrees ]]The interval of time is,take on [30 degrees, 90 degrees ]]An interval.
2.4) calculating the final matched bifurcation point pair set Afinal；
Because the correct point pairs are closely associated, the bifurcation point matching problem is converted into the node clustering problem of the graph G, and then the solution is carried out by utilizing a characteristic vector method, and finally, all right-matching bifurcation point pair sets are obtained and are marked as Afinal. The specific steps of solving the algorithm are as follows:
2.4.1) establishing a final matched bifurcation point pair set AfinalAnd A isfinalInitializing to an empty set;
2.4.2) decomposing the eigenvalue of the matrix M to find the eigenvector x corresponding to the maximum eigenvalue of M*；
2.4.3) for x*The elements in the sequence are sorted from big to small to obtain the sequence x ', and the elements of the sequence x' from front to back are arranged in the sequence x*The sequence numbers in (1) form a sequence L;
2.4.4) determine L: if L is null, then output the current AfinalEnding the solving algorithm; if L is not empty, entering step 2.4.5);
2.4.5) takes the first value L (1) in the sequence L and decides: if x*(L (1)) < ε, the current A is outputfinalAnd ending the solving algorithm, wherein the value of epsilon is [0.0000001, 0.01 ]]An interval; otherwise, entering step 2.4.6);
2.4.6) if AcandThe L (1) th branch point pair in (A)finalIf any pair of the branch points contains the same branch point, deleting L (1) from L, and returning to the step 2.4.4); otherwise, go to step 2.4.7);
2.4.7) A is treatedcandThe L (1) th branch point pair in (A) is added into the set AfinalIn (3), L (1) is deleted from L, and the step 2.4.4) is returned again.
3) Updating the central line in each coronary artery image to obtain a deleted central line segment set; matching the centerline segments by using the updated centerline and the finally matched bifurcation point pair set in the step 2); the method comprises the following specific steps:
a vessel segment is defined as a vessel portion sandwiched between two bifurcation points of a coronary artery, or a vessel portion sandwiched between a bifurcation point and a starting point (starting point refers to an exit of a coronary artery connected to an aorta) or an ending point (ending point refers to an ending point of a peripheral coronary artery) of a coronary artery. The centerline segment of the coronary artery is defined as the centerline of the vessel segment. The following definitions are made for three centerline segments connected to a bifurcation point: if the other end of the centerline segment is closer to the origin of the coronary artery relative to the bifurcation point, the segment is called the "trunk segment" of the bifurcation point; if the other end of the centerline segment is further from the origin of the coronary artery relative to the bifurcation, the segment is called a "branch segment" of the bifurcation. Each bifurcation point has one trunk segment and two branch segments. The method comprises the following specific steps:
3.1) updating the central line in each coronary artery image to obtain a deleted central line segment set
In many cases, there is a topological inconsistency between the centerlines of the two coronary arteries to be registered, and after the matching bifurcation point of step 2), there still exist some unmatched bifurcation points, i.e. the bifurcation points unique to the centerlines of the coronary arteries in one image. The specific treatment method comprises the following steps:
3.1.1) obtaining CsOr CtRespectively calculating the direction vectors (taking the direction vector pointing from the near end of the heart to the far end of the heart) of two branch centerline segments and a trunk centerline segment corresponding to the specific bifurcation point at the bifurcation point, deleting the branch centerline segment with a larger included angle between the direction vector and the direction vector of the trunk centerline segment, and deleting the specific bifurcation point at the same time.
3.1.2) from CsAnd CtAnd continuously searching for a unique bifurcation point from the remaining bifurcation points and judging: if the unique bifurcation point exists, returning to the step 3.1.1) again; if no specific bifurcation point exists, ending the centerline updating process to obtain the coronary artery centerline of the updated source image as C'sAnd the coronary artery center line of the updated target image is recorded as C'tAnd recording the Set of all deleted centerline segments as Setseg_del；
The procedure for the centerline update described above is illustrated as follows:
fig. 3 shows a schematic diagram of the coronary artery centerline in two coronary artery images with different topologies in this embodiment, and fig. 3(a) shows the coronary artery centerline C of the source imagesFIG. 3(b) is a coronary artery centerline C of the object imaget；
As shown in FIG. 3, CsWith a characteristic bifurcation point bf2And C istNone. The vessel segment matching algorithm of the present invention will remove CsCharacteristic bifurcation point bf of2And a centerline segment c connected thereto5And fragment c is inserted3And c4Merge into a new fragment, thus CsIs given and CtA consistent new topology. Note that c is removed here5And not c4Is due to the fact that5In contrast, c4And c3The included angle of the direction vector of (a) is smaller.
Fig. 4 is an exemplary diagram of a coronary artery centerline with inconsistent topology in practical application of the present invention. FIG. 4(a) is a coronary artery center line C of a source imagesFIG. 4(b) is a coronary artery centerline C of the object imaget. In the practical case shown in fig. 4, C is indicated by a dotted linesAre sequentially removed from the center line of the source image.
3.2) matching of centerline segments;
from C 'obtained in step 3.1)'sAnd C'tAnd the final matched bifurcation point pair set A obtained in the step 2)finalCenterline segments that satisfy one of any two of the following conditions are considered a match: a) two centerline segments sandwiched between two pairs of matched bifurcation point pairs; b) one end is a coronary artery starting point or end point, the other end is a matched bifurcation point pair, and the direction included angle of two bifurcation points of the bifurcation point pair is smaller than a set included angle threshold value (the value range of the included angle threshold value of the invention is 10 degrees and 60 degrees)]Interval) of two centersA line segment. At this time, pairs of centerline segments are obtained, and the Set of matched centerline segment pairs is referred to as Setseg_pair；
3.3) fine registration of centerline segments;
after the correspondence relationship of the blood vessel segments is clarified, Set is requiredseg_pairEach pair of centerline segments in (a) is fine registered separately. Wherein Set is to be Setseg_pairThe segment on the center line belonging to the source image is marked as Segs，Setseg_pairThe center line segment belonging to the target image is marked as Segt. For SegsAnd SegtPoint sampling, Seg, is performed separatelysIs a sampling interval ofs，SegtIs a sampling interval oft，ΔsAnd ΔtMay or may not be equal (suggest Δ)t/ΔsTake on the value of [1, 20]Interval). Let SegsAnd SegtRespectively have NsAnd NtSampling points, defining a discrete transformation model among a point set by using a series of states as follows:
S：＝{S(i)＝j，i∈{0，1，...，Nt}，j∈{0，1，...，Ns}}，
wherein S (i) is SegtThe state of the ith sample point of (1). S (i) ═ j denotes SegtThe ith sampling point of (1) and the SegsCorresponds to the jth sample point of (a). In the invention, an objective function is designed and maximized to solve parameters of the model S, namely values of i and j in S. The objective function is composed of two parts of image similarity and geometric similarity, and the overall objective function is as follows:
where Sim1 represents image similarity, Sim2 represents geometric similarity, ω1And ω2The weights occupied by Sim1 and Sim2, respectively. Omega1And ω2All values of (A) are [0, 1 ]]Interval, and ω1And omega2The sum of (1).
The image similarity Sim1 is measured by using a feature descriptor (for example, a feature descriptor at a bifurcation point, or other feature descriptors may be used), and the mathematical expression is as follows:
wherein,is SegtA descriptor of the ith sample point of (a),is Segs(ii) a descriptor of the sampling point(s). In order to avoid the central line from folding or over-telescoping, geometric constraint needs to be applied to the central line, and the corresponding geometric similarity expression is as follows:
because the state S (i) is only related to the state S (i-1) and meets Markov property, the solution problem of the transformation S can be regarded as a special hidden Markov model, and the solution problem is solved by using a Viterbi algorithm to obtain values of i and j in the S, so that a series of pairs of (i, j) points can be obtained. Will Setseg_pairAnd merging the { i, j } point pairs obtained by fine registration of each pair of centerline segments to form a point pair set B. So far, the corresponding relation of the sampling points on the paired segments is clear. Fig. 5(a) and (b) show the registration results of two pairs of centerline segments, respectively. The samples in the figure are denoted by the symbol "+", the successfully matched samples are denoted by the symbol "o" and are given a numerical label,are not numbered becauseIs greater thanAre partially numbered inThere is no corresponding point becauseIs longer.
4) Further segmenting and registering the centerline segments by using the centerline segment set deleted in the step 3) to obtain a final registration result of the centerline of the coronary artery;
step 3.1) deleted centerline segment Setseg_delThese centerline segments and the vessel segments corresponding to them are considered as the segments missing from the initial segmentation and extraction method; this step is used to process the part of the input that is inconsistent in the topology of the two coronary artery centerlines, Setseg_delMissing centerline segments in (1). Firstly, segmentation and extraction are carried out on the missing vessel segment and the central line segment, and then registration is carried out on the newly extracted central line segment. The method comprises the following specific steps:
4.1) further segmenting and extracting the missed blood vessel segments and the central line segments;
assuming the coronary centerline C of the original source imagesHaving a centre line segment CSs(i.e., coronary vessel V)sWith vessel segments VSs) And the coronary artery center line C of the target imagetHas missed the corresponding segment CSt(i.e. V)tOmit the corresponding vessel segment VSt) This step attempts to get from CtAnd VtIn the imageSegmentation of missing vessel segments VStAnd extracting the missing centerline segment CSt. Firstly, a continuous space transformation model (such as a thin plate spline model) is calculated by mathematical interpolation by using the point pair set B finally obtained in the step 3), and is recorded as T. Using T pairs of VSsPerforming a spatial transformation to obtain a mask VS in the other imagea. At VSsAnd VSaRespectively define the interested regions, denoted as VOIsAnd VOIt. One region of interest that may be employed is a block of a cuboid image on the image, each face of the cuboid being parallel to the plane of the coordinate axis, containing a mask and being spaced from the mask boundary in a direction parallel to the coordinate axis by a distance (value 0, 10)]Interval in millimeters). Registering two interested regions based on gray level to obtain space transformation model, and recording as TinterWill TinterIs applied to VSsTo obtainWill be provided withAs an initialization, the vessel branch VS is segmented from the target image using an image segmentation method (e.g. level set method)t。
For Setseg_delAnd (3) repeating the step to segment the blood vessels corresponding to all the central line segments to obtain corresponding newly segmented blood vessel branches, wherein the synchronous step 1.2) can be used for extracting the central line segment of each new blood vessel branch by utilizing manual labeling or semi-automatic and full-automatic algorithms.
Fig. 6 is an example graph showing the segmentation result of the more complete coronary artery blood vessel obtained by the step 4.1), and fig. 6(a) and fig. 6(b) are a left coronary artery blood vessel and a right coronary artery blood vessel respectively, wherein a black continuous surface is the initially input coronary artery blood vessel, and a discrete point surface is the blood vessel branch obtained by the further segmentation of the step 4.1).
4.2) for the newly extracted centerline segment of step 4.1), utilize step 3) The method carries out center line segment registration to the obtained point pairs to obtain a series of new sampling point pairs. Expanding the point pair set B by using the points to obtain a new point pair set Bfurther. By means of BfurtherPerforming mathematical interpolation to calculate continuous space transformation model, and recording as Tfinal. By TfinalCarrying out spatial transformation on a source image to obtain a transformed image aligned with a target image; by TfinalFor coronary artery blood vessel V in source imagesPerforming spatial transformation to obtain a sum VtAligned blood vessel transformation Va(ii) a By TfinalFor the coronary artery central line C in the source imagesPerforming spatial transformation to obtain CtAligned transformation center line Ca。BfurtherTransforming image VaAnd CaWhich is the final coronary artery registration result obtained by the present invention.
Fig. 7 and 8 show two examples of the registration results of the present invention, in which fig. 7(a) and 8(a) are the preliminary registration results obtained through step 2) and step 3), and fig. 7(b) and 8(b) are the final registration results processed through step 4). Coronary artery central line C in target image in the figuretIs represented by ". multidot." line type, CsAnd CsThe centerline obtained by the spatial transformation using T is represented by the "-" line, and the centerline that is successfully registered (after registration, the distance between the two centerlines is less than 0.5 mm) is represented by the solid line. It can be seen from fig. 7 and 8 that step 4) allows more centerline segments to be registered.
Claims (1)
1. A method for automatic registration of coronary arteries, characterized in that it comprises the following steps:
1) acquiring coronary artery images, performing blood vessel segmentation on each coronary artery image and extracting a corresponding central line; the method comprises the following specific steps:
1.1) acquiring a coronary artery image;
acquiring a pair of computed tomography angiography images which are performed aiming at one coronary artery, and randomly selecting one of the two images as a source image and the other image as a target image;
1.2) segmenting coronary artery blood vessels for each image obtained in the step 1.1), and extracting a central line corresponding to each coronary artery blood vessel; the coronary artery blood vessel in the source image is recorded as VsThe center line is marked as Cs(ii) a Coronary artery vessels in the target image are registered as VtThe center line is marked as Ct；
2) Extracting bifurcation point characteristics from the central line in each coronary artery image in the step 1) and matching to obtain a final matched bifurcation point pair set; the method comprises the following specific steps:
2.1) extracting the features of the bifurcation points;
let BsRepresenting the whole n of the coronary artery central lines in the source imagesA point set of points of bifurcation, BtRepresenting the whole of the coronary centerline n in the target imagetA point set consisting of multiple bifurcation points; extracting a corresponding feature descriptor for a bifurcation point of each coronary artery centerline, and representing the descriptor in a vector form; wherein the descriptor vector of the ith bifurcation point of the source image is taken asThe descriptor vector of the ith bifurcation point in the target image is taken as
2.2) obtaining a candidate matching bifurcation point pair set Acand；
Calculation of BtEach branch point b intAnd BsThe Euclidean distance of the descriptor vectors of all the bifurcation points is calculated, and B corresponding to the minimum Euclidean distance in the result is calculatedsThe bifurcation point in (1) is denoted by btCorresponding bifurcation point bsIn total, n is obtainedtA point pair of bifurcations, ntThe set of the branch point pairs is marked as a candidate matching branch point pair set Acand；
2.3) computing the set AcandThe attribute matrix M of (2);
first, wait AcandA graph G (V, E, M) is established, wherein each node in the set of nodes V of the graph corresponds to AcandEach side in the side set E of the graph corresponds to the relationship between every two bifurcation points, the off-diagonal element M (i, j) of the attribute matrix M reflects the compatibility between the ith bifurcation point pair and the jth bifurcation point pair, and the diagonal element M (i, i) of the attribute matrix M represents the descriptor sub-vector similarity of the two bifurcation points in the ith bifurcation point pair; the expression of M is as follows:
wherein d isiRepresenting two bifurcation points m in the ith bifurcation point pairi1And mi2Euclidean distance of thetaiIs mi1And mi2The relative angle of the directions is such that,is a handle mi2M when the corresponding vector is taken as the polar axisi1The polar angle of (d); THdist、THθAndthreshold values, | θ, representing lower limits of ratio (i, j), respectivelyi-θjThreshold sum of | upper boundA threshold value of an upper limit; THdistTake on the value of [0.4, 0.9]Interval, THθTake on [30 degrees, 90 degrees ]]The interval of time is,take on [30 degrees, 90 degrees ]]An interval;
2.4) calculating the final matched bifurcation point pair set Afinal(ii) a The method comprises the following specific steps:
2.4.1) establishing a final matched bifurcation point pair set AfinalAnd A isfinalInitializing to an empty set;
2.4.2) decomposing the eigenvalue of the matrix M to find the eigenvector x corresponding to the maximum eigenvalue of M*；
2.4.3) for x*The elements in the sequence are sorted from big to small to obtain the sequence x ', and the elements of the sequence x' from front to back are arranged in the sequence x*The sequence numbers in (1) form a sequence L;
2.4.4) determine L: if L is null, then output the current AfinalAnd ending the solution; if L is not empty, entering step 2.4.5);
2.4.5) takes the first value L (1) in the sequence L and decides: if x*(L (1)) < ε, the current A is outputfinalAnd ending the solution, wherein the value of epsilon is [0.0000001, 0.01 ]]An interval; otherwise, entering step 2.4.6);
2.4.6) if AcandThe L (1) th branch point pair in (A)finalIf any pair of the branch points contains the same branch point, deleting L (1) from L, and returning to the step 2.4.4); otherwise, go to step 2.4.7);
2.4.7) A is treatedcandThe L (1) th branch point pair in (A) is added into the set AfinalDeleting L (1) from L, and returning to the step 2.4.4);
3) updating the central line in each coronary artery image to obtain a deleted central line segment set; matching the centerline segments by using the updated centerline and the finally matched bifurcation point pair set in the step 2); the method comprises the following specific steps:
3.1) updating the central line in each coronary artery image to obtain a deleted central line segment set; the method comprises the following specific steps:
3.1.1) obtaining CsOr CtRespectively calculating the direction vectors of two branch centerline segments and a trunk centerline segment corresponding to the special bifurcation point at the bifurcation point, deleting the branch centerline segment with a larger included angle between the direction vectors and the direction vectors of the trunk centerline segments, and deleting the special bifurcation point;
3.1.2) from CsAnd CtAnd continuously searching for a unique bifurcation point from the remaining bifurcation points and judging: if the unique bifurcation point exists, returning to the step 3.1.1) again; if no specific bifurcation point exists, ending the centerline updating process to obtain the coronary artery centerline of the updated source image as C'sAnd the coronary artery center line of the updated target image is recorded as C'tAnd recording the Set of all deleted centerline segments as Setseg_del；
3.2) matching of centerline segments;
from C 'obtained in step 3.1)'sAnd C'tAnd the final matched bifurcation point pair set A obtained in the step 2)finalCenterline segments that satisfy one of any two of the following conditions are considered a match: a) two centerline segments sandwiched between two pairs of matched bifurcation point pairs; b) one end is a coronary artery starting point or end point, the other end is a matched bifurcation point pair, and the direction included angle of two bifurcation points of the bifurcation point pair is smaller than two central line segments of a set included angle threshold value; the Set of matched centerline segment pairs is denoted as Setseg_pair；
3.3) fine registration of centerline segments;
will Setseg_pairThe segment on the center line belonging to the source image is marked as Segs，Setseg_pairThe center line segment belonging to the target image is marked as Segt(ii) a For SegsAnd SegtPoint sampling, Seg, is performed separatelysIs a sampling interval ofs，SegtIs a sampling interval oft；
Let SegsAnd SegtRespectively have NsAnd NtSampling points, defining a discrete transformation model between a set of points as follows:
S：＝{S(i)＝j，i∈{0，1，…，Nt}，j∈{0，1，…，Ns}}，
wherein S (i) is SegtS (i) ═ j denotes the state of the ith sample point of (g)tThe ith sampling point of (1) and the SegsCorresponds to the jth sample point of (a);
the objective function is:
where sim1 represents image similarity, sim2 represents geometric similarity, ω1And ω2The weights, ω, occupied by Sim1 and Sim2, respectively1And ω2All values of (A) are [0, 1 ]]Interval, and ω1And omega2The sum of (a) and (b) is 1;
the image similarity Sim1 is measured by using a feature descriptor, and the mathematical expression is as follows:
wherein,is SegtA descriptor of the ith sample point of (a),is Segs(ii) a descriptor of the sample point of (a);
the geometric similarity Sim2 is expressed as:
solving the values of i and j in the S to obtain a series of (i, j) point pairs; will Setseg_pairMerging the { i, j } point pairs obtained by fine registration of each pair of central line segments to form a point pair set B;
4) further segmenting and registering the centerline segments by using the centerline segment set deleted in the step 3) to obtain a final registration result of the centerline of the coronary artery; the method comprises the following specific steps:
4.1) assuming the coronary centerline C of the source imagesHaving a centre line segment CSsI.e. coronary vessels VsWith vessel segments VSsAnd the coronary artery center line C of the target imagetHas missed the corresponding segment CStI.e. VtOmit the corresponding vessel segment VStCarrying out mathematical interpolation on the point pair set B obtained in the step 3) to calculate a continuous space transformation model, and recording the model as T; using T pairs of VSsPerforming a spatial transformation to obtain a mask VS in the other imageaAt VSSAnd VSaRespectively define the interested regions, denoted as VOIsAnd VOIt(ii) a Registering two interested regions based on gray level to obtain space transformation model, and recording as TinterWill TinterIs applied to VSsTo obtainWill be provided withAs an initialization, the vessel branch VS is segmented from the target image by means of an image segmentation methodt；
For Setseg_delRepeating the steps for the blood vessels corresponding to all the central line segments to obtain the corresponding blood vesselsNewly segmenting the blood vessel branches, and then repeating the step 1.2) to obtain central line segments of all the newly segmented blood vessel branches;
4.2) for the centerline segment newly extracted in the step 4.1), repeating the step 3) to carry out centerline segment registration on the centerline segment, obtaining a series of new sampling point pairs and adding the new sampling point pairs into the set B to obtain a new point pair set Bfurther(ii) a By means of BfurtherPerforming mathematical interpolation to calculate continuous space transformation model, and recording as Tfinal(ii) a By TfinalCarrying out spatial transformation on a source image to obtain a transformed image aligned with a target image; by TfinalFor coronary artery blood vessel V in source imageSPerforming spatial transformation to obtain a sum VtAligned blood vessel transformation Va(ii) a By TfinalFor the coronary artery central line C in the source imagesPerforming spatial transformation to obtain CtAligned transformation center line Ca；BfurtherTransforming image VaAnd CaI.e. the final result of the registration of the centerline of the coronary artery.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810643770.6A CN108898626B (en) | 2018-06-21 | 2018-06-21 | A kind of autoegistration method coronarius |
| PCT/CN2018/116965 WO2019242227A1 (en) | 2018-06-21 | 2018-11-22 | Automatic registration method for coronary arteries |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810643770.6A CN108898626B (en) | 2018-06-21 | 2018-06-21 | A kind of autoegistration method coronarius |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108898626A CN108898626A (en) | 2018-11-27 |
| CN108898626B true CN108898626B (en) | 2019-09-27 |
Family
ID=64345225
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810643770.6A Active CN108898626B (en) | 2018-06-21 | 2018-06-21 | A kind of autoegistration method coronarius |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN108898626B (en) |
| WO (1) | WO2019242227A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111932497A (en) * | 2020-06-30 | 2020-11-13 | 数坤（北京）网络科技有限公司 | Coronary artery identification method and device |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109745120B (en) * | 2018-12-24 | 2020-07-31 | 罗雄彪 | Image registration conversion parameter optimization method and system |
| US10813612B2 (en) | 2019-01-25 | 2020-10-27 | Cleerly, Inc. | Systems and method of characterizing high risk plaques |
| CN109978829B (en) * | 2019-02-26 | 2021-09-28 | 深圳市华汉伟业科技有限公司 | Detection method and system for object to be detected |
| CN109993730B (en) * | 2019-03-20 | 2021-03-30 | 北京理工大学 | 3D/2D blood vessel registration method and device |
| CN110197495B (en) * | 2019-05-30 | 2021-03-09 | 数坤（北京）网络科技有限公司 | Adjusting method and device for blood vessel extraction |
| JP2023509514A (en) | 2020-01-07 | 2023-03-08 | クリールリー、 インコーポレーテッド | Systems, Methods, and Devices for Medical Image Analysis, Diagnosis, Severity Classification, Decision Making, and/or Disease Tracking |
| US20220392065A1 (en) | 2020-01-07 | 2022-12-08 | Cleerly, Inc. | Systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking |
| US11969280B2 (en) | 2020-01-07 | 2024-04-30 | Cleerly, Inc. | Systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking |
| CN111242958B (en) * | 2020-01-15 | 2022-04-08 | 浙江工业大学 | Carotid artery cascade learning segmentation method based on structural feature optimization |
| US20230289963A1 (en) | 2022-03-10 | 2023-09-14 | Cleerly, Inc. | Systems, devices, and methods for non-invasive image-based plaque analysis and risk determination |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104574413A (en) * | 2015-01-22 | 2015-04-29 | 深圳大学 | Blood vessel bifurcation extracting method and system of lung CT picture |
| CN105741251A (en) * | 2016-03-17 | 2016-07-06 | 中南大学 | Blood vessel segmentation method for liver CTA sequence image |
| CN105761254A (en) * | 2016-02-04 | 2016-07-13 | 浙江工商大学 | Image feature based eyeground image registering method |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7397935B2 (en) * | 2004-05-10 | 2008-07-08 | Mediguide Ltd. | Method for segmentation of IVUS image sequences |
| CN101763642B (en) * | 2009-12-31 | 2011-09-14 | 华中科技大学 | Matching method for three-dimensional coronary angiography reconstruction |
-
2018
- 2018-06-21 CN CN201810643770.6A patent/CN108898626B/en active Active
- 2018-11-22 WO PCT/CN2018/116965 patent/WO2019242227A1/en active Application Filing
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104574413A (en) * | 2015-01-22 | 2015-04-29 | 深圳大学 | Blood vessel bifurcation extracting method and system of lung CT picture |
| CN105761254A (en) * | 2016-02-04 | 2016-07-13 | 浙江工商大学 | Image feature based eyeground image registering method |
| CN105741251A (en) * | 2016-03-17 | 2016-07-06 | 中南大学 | Blood vessel segmentation method for liver CTA sequence image |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111932497A (en) * | 2020-06-30 | 2020-11-13 | 数坤（北京）网络科技有限公司 | Coronary artery identification method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108898626A (en) | 2018-11-27 |
| WO2019242227A1 (en) | 2019-12-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108898626B (en) | A kind of autoegistration method coronarius | |
| Tobon-Gomez et al. | Benchmark for algorithms segmenting the left atrium from 3D CT and MRI datasets | |
| CN112884826B (en) | Method and device for extracting center line of blood vessel | |
| US7996060B2 (en) | Apparatus, method, and computer software product for registration of images of an organ using anatomical features outside the organ | |
| CN108961273B (en) | Method and system for segmenting pulmonary artery and pulmonary vein from CT image | |
| KR102050649B1 (en) | Method for extracting vascular structure in 2d x-ray angiogram, computer readable medium and apparatus for performing the method | |
| CN109478327B (en) | Method for automatic detection of systemic arteries in Computed Tomography Angiography (CTA) of arbitrary field of view | |
| Wu et al. | Fast catheter segmentation from echocardiographic sequences based on segmentation from corresponding X-ray fluoroscopy for cardiac catheterization interventions | |
| Hoffmann et al. | Electrophysiology catheter detection and reconstruction from two views in fluoroscopic images | |
| CN102831614B (en) | Sequential medical image quick segmentation method based on interactive dictionary migration | |
| Rivas-Villar et al. | Color fundus image registration using a learning-based domain-specific landmark detection methodology | |
| US20130331687A1 (en) | Combined Cardiac and Respiratory Motion Compensation for Atrial Fibrillation Ablation Procedures | |
| CN113538471B (en) | Plaque segmentation method, plaque segmentation device, computer equipment and storage medium | |
| CN114332013B (en) | CT image target lung segment identification method based on pulmonary artery tree classification | |
| CN114081625B (en) | Navigation path planning method, system and readable storage medium | |
| JP2024059614A (en) | Method and device for processing blood vessel video based on user input | |
| Zhang et al. | Pathological airway segmentation with cascaded neural networks for bronchoscopic navigation | |
| CN110853020B (en) | Method for measuring retinal vascular network similarity based on topological structure and map | |
| CN116612166A (en) | Registration fusion algorithm for multi-mode images | |
| EP4082444A1 (en) | Automatic frame selection for 3d model construction | |
| CN112102352B (en) | Coronary artery motion tracking method and device for DSA image sequence | |
| Yang et al. | Scale-aware auto-context-guided Fetal US segmentation with structured random forests | |
| Brieva et al. | Coronary extraction and stenosis quantification in X-ray angiographic imaging | |
| CN114119688A (en) | Single-mode medical image registration method before and after coronary angiography based on deep learning | |
| CN112669370B (en) | Coronary artery radius calculation method, terminal and storage medium |
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 |