CN109035194B - Blood vessel extraction method and device - Google Patents
Blood vessel extraction method and device Download PDFInfo
- Publication number
- CN109035194B CN109035194B CN201810153950.6A CN201810153950A CN109035194B CN 109035194 B CN109035194 B CN 109035194B CN 201810153950 A CN201810153950 A CN 201810153950A CN 109035194 B CN109035194 B CN 109035194B
- Authority
- CN
- China
- Prior art keywords
- pixel point
- point
- pixel
- cur
- aorta
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
- G06T7/74—Determining position or orientation of objects or cameras using feature-based methods involving reference images or patches
-
- 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
Abstract
The application provides a blood vessel extraction method and a device, which relate to the technical field of image processing, and the method comprises the following steps: determining an aorta region in a medical image to be detected based on a designated pixel point, wherein the designated pixel point is positioned on an aorta; extracting a central axis of the aorta region; determining an end point of the central axis; and taking the end point as a starting point, and searching from the starting point by adopting an improved minimum path algorithm with backtracking to obtain a blood vessel region in the medical image to be detected, wherein in the improved minimum path algorithm with backtracking, an energy value of a pixel point is calculated based on a saliency characteristic value of the pixel point, and the saliency characteristic value of the pixel point is positively correlated with a gray value of the pixel point. By applying the method, the blood vessels in the medical image to be detected can be efficiently, completely and accurately extracted.
Description
Technical Field
The present application relates to the field of image processing technologies, and in particular, to a method and an apparatus for extracting blood vessels.
Background
The vascular diseases have the characteristics of high mortality, high disability rate, high medical risk and the like, and are one of the diseases seriously harming human health. In modern medicine, diagnosis and treatment of vascular diseases are usually performed by means of medical images, such as CT images, magnetic resonance images, etc., wherein the extraction of blood vessels on the medical images is a key step in the diagnosis and treatment process.
At present, a central line-based blood vessel extraction method is applied more, and the central line-based blood vessel extraction method is divided into a direct central line tracking method and a minimum path algorithm, wherein the direct central line tracking method needs to manually designate a pixel point as a root node of a blood vessel tree, and the principle of the method is that the blood vessel direction is estimated based on a model and image information to predict the position of a central line, and in order to increase robustness, the blood vessel extraction is generally carried out by combining filtering technologies such as Kalman filtering, Frangi and filtering; the minimum path algorithm needs to manually designate two pixel points as an initial point and an end point respectively, and aims to search for the shortest path between the two points.
However, in the direct centerline tracking method, only one branch can be processed at a time, and the blood vessel has a complicated structure and has a plurality of branches, so that the efficiency of extracting the blood vessel using the direct centerline tracking method is low; in the minimum path algorithm, because two pixel points need to be specified manually, manual interaction is more, which results in poor user experience, and meanwhile, when a search path is longer, the search cost is gradually accumulated, so that the efficiency of the whole algorithm is reduced, and in a complex situation, for example, in a situation with more branches, the minimum path algorithm also has a greater limitation.
Disclosure of Invention
In view of this, the present application provides a blood vessel extraction method and apparatus, so as to efficiently, completely and accurately extract a blood vessel in a medical image to be detected.
Specifically, the method is realized through the following technical scheme:
according to a first aspect of embodiments of the present application, there is provided a blood vessel extraction method, the method including:
determining an aorta region in a medical image to be detected based on a designated pixel point, wherein the designated pixel point is positioned on an aorta;
extracting a central axis of the aorta region;
determining an end point of the central axis;
and taking the end point as a starting point, and searching from the starting point by adopting an improved minimum path algorithm with backtracking to obtain a blood vessel region in the medical image to be detected, wherein in the improved minimum path algorithm with backtracking, an energy value of a pixel point is calculated based on a saliency characteristic value of the pixel point, and the saliency characteristic value of the pixel point is positively correlated with a gray value of the pixel point.
Optionally, the determining the aorta region in the medical image to be detected based on a specific pixel point includes:
taking a designated pixel point as an initial seed point;
and taking the initial seed point as a starting point, and performing region growth from the starting point by adopting a region growth algorithm to obtain an aorta region in the medical image to be detected.
Optionally, the determining the end point of the central axis includes:
calculating the number of connected pixels on the central axis;
and determining the pixel points with the connection number of 1 as the end points of the central axis.
Optionally, the saliency characteristic value of the pixel point is calculated by the following method:
calculating the saliency characteristic value of the pixel point by adopting a first preset formula, wherein the first preset formula is as follows:
M＝f(p_cur)*c(θ)；
wherein, M represents the saliency characteristic value of the pixel point, c (theta) represents the saliency of the pixel point in the theta direction, p _ cur represents the gray value of the pixel point, and f (p _ cur) represents a monotonically increasing nonlinear function related to the gray value of the pixel point.
Optionally, in the improved minimum path algorithm with backtracking, a second preset formula is adopted to calculate the accumulated energy of the pixel points, where the second preset formula is as follows:
P_cur'＝P_cur-P_bk；
the P _ cur' represents the final accumulated energy of the pixel point, the P _ cur represents the initial accumulated energy of the pixel point, and the P _ bk represents the accumulated energy of the pixel point reached after the set step number is traced from the pixel point.
According to a second aspect of embodiments of the present application, there is provided a blood vessel extraction device, the device comprising:
the aorta determining module is used for determining an aorta region in the medical image to be detected based on a specified pixel point, and the specified pixel point is positioned on the aorta;
the central axis extraction module is used for extracting a central axis of the aorta area;
an endpoint determination module for determining an endpoint of the central axis;
and the searching module is used for searching from the starting point by adopting an improved minimum path algorithm with backtracking to obtain a blood vessel region in the medical image to be detected, wherein in the improved minimum path algorithm with backtracking, the energy value of the pixel point is calculated based on the saliency characteristic value of the pixel point, and the saliency characteristic value of the pixel point is positively correlated with the gray value of the pixel point.
Optionally, the aorta determination module comprises:
the seed point determining submodule is used for taking a specified pixel point as an initial seed point;
and the region growing sub-module is used for taking the initial seed point as a starting point and performing region growing from the starting point by adopting a region growing algorithm to obtain the aorta region in the medical image to be detected.
Optionally, the endpoint determining module includes:
the connection number calculation submodule is used for calculating the connection number of the pixel points on the central axis;
and the determining submodule is used for determining the pixel points with the connection number of 1 as the end points of the central axis.
Optionally, the search module includes:
the energy value calculating operator module is used for calculating the saliency characteristic value of the pixel point by adopting a first preset formula, and the first preset formula is as follows:
M＝f(p_cur)*c(θ)；
wherein, M represents the saliency characteristic value of the pixel point, c (theta) represents the saliency of the pixel point in the theta direction, p _ cur represents the gray value of the pixel point, and f (p _ cur) represents a monotonically increasing nonlinear function related to the gray value of the pixel point.
Optionally, the search module includes:
an accumulated energy calculation submodule, configured to calculate, in the improved minimum path algorithm with backtracking, accumulated energy of a pixel point by using a second preset formula, where the second preset formula is as follows:
P_cur'＝P_cur-P_bk；
the P _ cur' represents the final accumulated energy of the pixel point, the P _ cur represents the initial accumulated energy of the pixel point, and the P _ bk represents the accumulated energy of the pixel point reached after the set step number is traced from the pixel point.
It can be seen from the above embodiments that an aorta region is determined in a medical image to be detected based on an appointed pixel point, the appointed pixel point is located on an aorta, a central axis of the aorta region is extracted, an end point of the central axis is further determined, and an improved minimum path algorithm with backtracking is adopted to search from the start point by taking the end point as a start point, so as to obtain a blood vessel region in the medical image to be detected, wherein in the improved minimum path algorithm with backtracking, an energy value of the pixel point is calculated based on a saliency characteristic value of the pixel point, and the saliency characteristic value of the pixel point is positively correlated with a gray value of the pixel point.
The process is realized only by the user designating one pixel point, so that manual operation is reduced, and user experience is improved; the aorta region determined based on one designated pixel point comprises the aorta trunk and partial vessel branches flowing from the aorta to other parenchymal viscera, so that the subsequently determined end points also comprise end points positioned on the partial branches, and the subsequent vessel extraction based on the end points can realize the minimum path search only once on the medical image to be detected, namely the complete vessel structure in the medical image to be detected can be extracted; because the minimum path search is carried out by adopting the improved minimum path algorithm with backtracking, in the algorithm, the convex characteristic value of the pixel point is positively correlated with the gray value of the pixel point, and the energy value of the pixel point is negatively correlated with the convex characteristic value of the pixel point, the energy value of the pixel point on the blood vessel with higher brightness can be smaller, the energy value of the pixel point on the non-blood vessel with lower brightness is larger, and in the minimum path search process, the pixel point on the blood vessel is easier to search, namely, the blood vessel is easier to extract, thereby improving the efficiency of blood vessel extraction on the image to be detected.
In summary, the blood vessel extraction method provided by the application can realize high-efficiency, complete and accurate extraction of the blood vessel in the medical image to be detected.
Drawings
FIG. 1 is a flow chart of an embodiment of a method for extracting blood vessels according to the present application;
FIG. 2 is an example of a medical image to be examined;
FIG. 3 is an example of an aortic region in the image illustrated in FIG. 2;
FIG. 4 is an illustration of a central axis of the aortic region illustrated in FIG. 3;
FIG. 5 is an example of an 8 neighborhood of pixel point x;
FIG. 6 is an example of an extracted vascular region;
FIG. 7 is another example of an extracted vascular region;
FIG. 8 is an example of an extracted intact vessel region;
FIG. 9 is a schematic view of a partial blood vessel;
fig. 10 is a hardware configuration diagram of an image processing device in which the blood vessel extraction device of the present application is located; .
Fig. 11 is a block diagram of an embodiment of a blood vessel extraction device according to the present application.
Detailed Description
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this application and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.
It is to be understood that although the terms first, second, third, etc. may be used herein to describe various information, such information should not be limited to these terms. These terms are only used to distinguish one type of information from another. For example, first information may also be referred to as second information, and similarly, second information may also be referred to as first information, without departing from the scope of the present application. The word "if" as used herein may be interpreted as "at … …" or "when … …" or "in response to a determination", depending on the context.
Vascular diseases have the characteristics of high mortality, high disability rate, high medical risk and the like, are gradually the first killers harmful to human health, and accurately and clearly extract the vascular structures in medical images, such as Computed Tomography (CT) images, nuclear magnetic resonance images and other images, which are important for diagnosis and treatment of vascular diseases.
Currently, there are a number of methods available for vessel extraction in medical images, such as region growing methods, matched filtering methods, and most commonly centerline-based methods. The region growing method has the advantages of simplicity, easiness in operation and the like, but due to the influence of factors such as more noise, uneven blood vessel brightness distribution and the like in a medical image, the blood vessel region extracted by the region growing method has the phenomena of holes, discontinuous edges and the like.
The matched filtering method can detect the distribution of blood vessels in different directions and blood vessels with different sizes, however, in order to obtain the final complete blood vessel structure, further connection processing is required, and the calculation amount of the connection processing is relatively large, so that the efficiency of extracting the blood vessels by using the matched filtering method is also low.
The blood vessel extraction method based on the center line is further divided into a direct center line tracking method and a minimum path method, wherein the direct center line tracking method needs to manually designate a pixel point as a root node of a blood vessel tree, the principle of the method is to estimate the blood vessel direction based on a model and image information to predict the position of the center line, and in order to increase robustness, the blood vessel extraction is generally performed by combining filtering technologies such as Kalman filtering and Frangi filtering, however, the direct center line tracking method can only process one branch within the same time, the blood vessel structure is complex, and the direct center line tracking method has a plurality of branches, so that the efficiency of blood vessel extraction by using the direct center line tracking method is low. The minimum path algorithm needs to manually designate two pixel points as an initial point and an end point respectively, and aims to search the shortest path between the two points.
Based on the description, the application provides a blood vessel extraction method, and in the blood vessel extraction method, an improved minimum path algorithm with backtracking is provided, so that the complete blood vessel structure in the medical image to be detected can be extracted only by providing one pixel point by a user, meanwhile, the extracted blood vessel structure has no hole, the edge is continuous and clear, subsequent processing is not needed, and the blood vessel extraction efficiency in the medical image to be detected is improved.
The following examples are provided to explain the blood vessel extraction method provided in the present application in detail.
Referring to fig. 1, a flow chart of an embodiment of the blood vessel extraction method of the present application is shown, and the method may include the following steps:
step 101: determining an aorta region in the medical image to be detected based on a designated pixel point, wherein the designated pixel point is located on the aorta.
Step 102: the central axis of the aortic region is extracted.
Step 103: the end points of the central axis are determined.
The above steps 101 to 103 are explained in detail as follows:
in the embodiment of the present application, the medical image to be detected may be a CT image, a magnetic resonance image, a fluoroscopy image, or the like, and taking the CT image as an example, as shown in fig. 2, it is an example of the medical image to be detected.
In the embodiment of the present application, a user may select a point on the aorta in the image illustrated in fig. 2, for example, the point P illustrated in fig. 2, and for convenience of description, the point selected by the user is referred to as a designated pixel point. Subsequently, the designated pixel point may be used as an initial seed point, and a region growing algorithm is used to perform region growing from the starting point by using the initial seed point as a starting point, so as to obtain an aorta region in the image illustrated in fig. 2, for example, as shown in fig. 3, which is an example of the aorta region in the image illustrated in fig. 2.
As shown in fig. 3, the obtained aortic region includes not only the main aorta but also small branches branching from the main aorta, which are partial blood vessel branches flowing from the aorta to parenchymal organs such as liver, spleen, and kidney.
Subsequently, for the aorta region illustrated in fig. 3, a refinement algorithm is further used to determine the central axis of the aorta region, for example, as shown in fig. 4, which is an example of the central axis of the aorta region illustrated in fig. 3, and regarding a specific process for determining the central axis of the aorta region illustrated in fig. 3 by using the refinement algorithm, those skilled in the art can refer to the related description in the prior art, and the detailed description of the present application is omitted here.
Subsequently, for the central axis illustrated in fig. 4, the connection number of each pixel point on the central axis may be calculated by using the following formula (one), and if the connection number of the pixel point is 1, the pixel point may be determined as an end point of the central axis, for example, the pixel point selected by a black box in fig. 4 represents the end point of the central axis, the central axis illustrated in fig. 4 has 5 end points, and for convenience of description, the 5 end points are respectively numbered as C1, C2, C3, C4, and C5.
In the above formula (one), NCRepresenting the number of connections of a pixel, f (x)k) Representing a pixel point xkK represents the k-th neighborhood of 8 neighborhoods, as shown in fig. 5, which is an example of 8 neighborhoods of the pixel point x.
Step 104: and taking the end point as a starting point, and searching from the starting point by adopting an improved minimum path algorithm with backtracking to obtain a blood vessel region in the medical image to be detected, wherein in the improved minimum path algorithm with backtracking, the energy value of a pixel point is calculated based on the saliency characteristic value of the pixel point, and the saliency characteristic value of the pixel point is positively correlated with the gray value of the pixel point.
In the embodiment of the present application, step 104 is performed on the 5 end points determined in step 103, for example, taking the end point C1 as an example, taking the end point C1 as a starting point, starting a search from the starting point by using the improved minimum path algorithm with backtracking, so as to obtain a blood vessel region as shown in fig. 6, then, on the basis of fig. 6, continuing to take the end point C2 as a starting point, starting a search from the starting point by using the improved minimum path algorithm with backtracking, obtaining a blood vessel region as shown in fig. 7, then, continuing to take the end points C3, C4, and C5 as starting points, and finally obtaining a complete blood vessel region in the image to be detected as shown in fig. 8.
It will be understood by those skilled in the art that the above-described sequential search sequence from the end point C1 to the end point C5 is merely an example, and the present application is not limited to which end point the search is initiated, and the search is performed on a per-end-point basis.
The improved minimum path with backtracking algorithm proposed in the present application is described as follows:
it will be appreciated by those skilled in the art that the minimum path algorithm in the prior art can implement the feature extraction by searching the path with the minimum accumulated energy between two preset points. The commonly used minimum path search algorithm comprises a Dijkstra algorithm, a fast approximation algorithm and the like, wherein the Dijkstra algorithm is taken as an example, firstly, the accumulated energy of a starting point is set to be 0, the accumulated energy of other pixel points except the starting point is set to be infinite, then, the searching is carried out in a neighborhood from the starting point, wherein for a two-dimensional image, the neighborhood can be an 8-point neighborhood, for a three-dimensional image, the neighborhood can be a 26-point neighborhood, in the searching process, the pixel points in the neighborhood are added into a priority queue, the pixel points in the priority queue are sorted by adopting a minimum heap structure, the pixel point with the minimum accumulated energy at present is selected as a new starting point, the searching is continued from the new starting point until the end point is searched, and the searching is stopped, so that the path with the minimum accumulated energy between the initial starting point and the end point can be obtained.
In the above described minimum path algorithm, firstly, two points, namely, the starting point and the end point, need to be specified in advance, which results in more manual interaction and poor user experience, secondly, since Dijkstra algorithm tends to find the shortest and closest curve structure between the starting point and the end point, a path with a short geodesic distance between the starting point and the end point will be mistakenly taken as the minimum path between the starting point and the end point, which is the short circuit problem in the existing minimum path algorithm, and finally, as the search process is continuously performed, the longer the characteristic point is from the starting point, the longer the length of the minimum path between the starting point and the characteristic point is, which means that the accumulated energy value of the remote characteristic point is accumulated by the capability values of more pixel points, so that the remote characteristic point is mistakenly determined as a non-characteristic point due to the larger accumulated energy value, and the true non-characteristic point is smaller due to the accumulated energy value, and the identified feature points, namely more and more non-feature points, are searched, namely the problem of cost accumulation in the existing minimum path algorithm.
Based on the above problem, a minimum path algorithm with backtracking is further proposed, specifically, in order to reduce the path cost of the far-end feature point, so that the far-end feature point can be quickly searched as a feature point closer to the starting point, and the accumulated energy of the far-end feature point is reduced, for example, the accumulated energy of the far-end feature point may be calculated by a second preset formula described below.
The second predetermined formula is: p _ cur ═ P _ cur-P _ bk;
wherein, P _ cur' represents the final accumulated energy of the pixel, P _ cur represents the initial accumulated energy of the pixel, and P _ bk represents the accumulated energy of the pixel reached after the step number is set from the backtracking of the pixel.
By adopting backtracking operation to reduce the accumulated energy of the far-end feature points, the search can be effectively carried out around the curve feature structure, so that the short circuit problem and the cost accumulation problem in the minimum path algorithm can be better overcome.
Further, based on the minimum path algorithm with backtracking, in order to achieve blood vessel extraction only according to one starting point without specifying an end point, a stopping search criterion may be preset, for example, a normalized average backtracking speed is used to set the stopping search criterion, and for example, the stopping search criterion is set by setting the number of iterations in the search process, and regarding the specific description of setting the stopping search criterion, those skilled in the art may refer to the related description in the prior art, and this application will not be described in detail herein.
It should be noted that, in the above-described minimum path algorithm with backtracking, the most critical point is the construction of the blood vessel energy function, in the prior art, the saliency characteristic value of the blood vessel is determined only by the saliency of the blood vessel, for example, the saliency characteristic value of the pixel point in the prior art can be calculated by the following formula (two):
M＝c(θ)2formula 2
In the above formula (two), M represents the saliency characteristic value of the pixel, c (θ) represents the saliency of the pixel in the θ direction, and the energy value of the pixel can be calculated by the following formula (three):
in the above formula (three), p represents the energy value of the pixel.
As can be seen from the above formula (iii), the larger the saliency characteristic value of a pixel point is, the lower the energy value of the pixel point is.
However, determining the saliency characteristic of a blood vessel based on the saliency of the blood vessel results in an energy value calculated from the saliency characteristic that does not distinguish between pixels on the blood vessel and pixels on non-blood vessels well, e.g., in the local blood vessel diagram illustrated in fig. 9, assuming that the P point is a pixel point on the blood vessel, and the P1 and the P2 are two points with the same distance from the P point, based on the definition of the saliency of the pixel point, it can be known that the saliency of P1 is the same as that of P2, and, on the line segment [ P1, P2], the convexity of two points symmetrical to the left and right is the same with the point P as the center, and then, the energy values of P1 and P2 calculated by the above formula (two) and the above formula (three) are the same, the energy values of two symmetrical points are also the same, it can be seen that the finally determined blood vessel region will cover the interval P1, P2, which is significantly larger than the real blood vessel region.
Based on this, the embodiment of the present application provides a new method for calculating a saliency characteristic value of a pixel point, and specifically, in consideration of a large difference between gray values of a pixel point on a blood vessel and a pixel point on a non-blood vessel, a gray value of a pixel point on the non-blood vessel is obviously smaller than a gray value of a pixel point on the blood vessel, so that the saliency characteristic value of the pixel point is determined based on the gray value of the pixel point and the saliency of the pixel point, for example, the saliency characteristic value of the pixel point can be calculated by using a first preset formula as follows:
a first preset formula: m ═ f (p _ cur) × c (θ);
in the first preset formula, p _ cur represents the gray value of the pixel point, f (p _ cur) represents a nonlinear function related to the gray value of the pixel point, and in order to better distinguish the blood vessel region from the non-blood vessel region, the convex characteristic value of the pixel point on the blood vessel is set to be greater than the convex characteristic value of the pixel point on the non-blood vessel, that is, the gray value of the pixel point is greater, the convex characteristic value is greater, and thus f (p _ cur) is a monotonically increasing function.
In one embodiment, f (p _ cur) can be (i) or (ii) as follows:
(i)
(ii)
in the above (i) or (ii), max _ p represents the maximum gray value in the medical image to be detected, and in the above (i), r _ cur represents the blood vessel radius corresponding to the estimated pixel point, wherein the blood vessel radius corresponding to the pixel point can be estimated by using the Ray-Casting method, and the specific process of estimating the blood vessel radius corresponding to the pixel point by using the Ray-Casting method can be referred to the related description in the prior art, which is not described in detail in this application.
It should be noted that (i) or (ii) above is only an example, and in practical applications, f (p _ cur) is a monotonically increasing nonlinear function related to the gray-scale value of the pixel, and the application does not limit the specific form of f (p _ cur).
It can be seen from the above embodiments that an aorta region is determined in a medical image to be detected based on an appointed pixel point, the appointed pixel point is located on an aorta, a central axis of the aorta region is extracted, an end point of the central axis is further determined, and an improved minimum path algorithm with backtracking is adopted to search from the start point by taking the end point as a start point, so as to obtain a blood vessel region in the medical image to be detected, wherein in the improved minimum path algorithm with backtracking, an energy value of the pixel point is calculated based on a saliency characteristic value of the pixel point, and the saliency characteristic value of the pixel point is positively correlated with a gray value of the pixel point.
The process is realized only by the user designating one pixel point, so that manual operation is reduced, and user experience is improved; the aorta region determined based on one designated pixel point comprises the aorta trunk and partial vessel branches flowing from the aorta to other parenchymal viscera, so that the subsequently determined end points also comprise end points positioned on the partial branches, and the subsequent vessel extraction based on the end points can realize the minimum path search only once on the medical image to be detected, namely the complete vessel structure in the medical image to be detected can be extracted; because the minimum path search is carried out by adopting the improved minimum path algorithm with backtracking, in the algorithm, the convex characteristic value of the pixel point is positively correlated with the gray value of the pixel point, and the energy value of the pixel point is negatively correlated with the convex characteristic value of the pixel point, the energy value of the pixel point on the blood vessel with higher brightness can be smaller, the energy value of the pixel point on the non-blood vessel with lower brightness is larger, and in the minimum path search process, the pixel point on the blood vessel is easier to search, namely, the blood vessel is easier to extract, thereby improving the efficiency of blood vessel extraction on the image to be detected.
In summary, the blood vessel extraction method provided by the application can realize high-efficiency, complete and accurate extraction of the blood vessel in the medical image to be detected.
Corresponding to the embodiment of the blood vessel extraction method, the application also provides an embodiment of the blood vessel extraction device.
The embodiment of the blood vessel extraction device can be applied to image processing equipment. The device embodiments may be implemented by software, or by hardware, or by a combination of hardware and software. Taking a software implementation as an example, as a device in a logical sense, the device is formed by reading a corresponding computer program instruction in a nonvolatile memory into an internal memory for operation through a processor of the image processing apparatus where the device is located. In terms of hardware, as shown in fig. 10, the present application is a hardware structure diagram of an image processing apparatus in which a blood vessel extraction device is located, except for the processor 101, the memory 102, the network interface 103, and the nonvolatile memory 104 shown in fig. 10, the image processing apparatus in which the device is located in the embodiment may also include other hardware according to an actual function of the image processing apparatus, which is not described again.
Referring to fig. 11, a block diagram of an embodiment of a blood vessel extraction device according to the present application is shown, the device including: an aorta determination module 111, a central axis extraction module 112, an end point determination module 113, and a search module 114.
The aorta determining module 111 may be configured to determine an aorta region in the medical image to be detected based on a designated pixel point, where the designated pixel point is located on the aorta;
a central axis extraction module 112, which may be configured to extract a central axis of the aortic region;
an end point determining module 113, which may be configured to determine an end point of the central axis;
the searching module 114 may be configured to start, with the endpoint as a starting point, searching from the starting point by using an improved minimum path algorithm with backtracking to obtain a blood vessel region in the medical image to be detected, where in the improved minimum path algorithm with backtracking, an energy value of a pixel point is calculated based on a saliency feature value of the pixel point, and the saliency feature value of the pixel point is positively correlated with a gray value of the pixel point.
In an embodiment, the aorta determination module 111 may comprise (not shown in fig. 11):
the seed point determining submodule is used for taking a specified pixel point as an initial seed point;
and the region growing sub-module is used for taking the initial seed point as a starting point and performing region growing from the starting point by adopting a region growing algorithm to obtain the aorta region in the medical image to be detected.
In an embodiment, the endpoint determination module 113 may include (not shown in fig. 11):
the connection number calculation submodule is used for calculating the connection number of the pixel points on the central axis;
and the determining submodule is used for determining the pixel points with the connection number of 1 as the end points of the central axis.
In one embodiment, the search module 114 may include (not shown in fig. 11):
the energy value calculating operator module is used for calculating the saliency characteristic value of the pixel point by adopting a first preset formula, and the first preset formula is as follows:
M＝f(p_cur)*c(θ)；
wherein, M represents the saliency characteristic value of the pixel point, c (theta) represents the saliency of the pixel point in the theta direction, p _ cur represents the gray value of the pixel point, and f (p _ cur) represents a monotonically increasing nonlinear function related to the gray value of the pixel point.
In one embodiment, the search module 114 may include (not shown in fig. 11):
an accumulated energy calculation submodule, configured to calculate, in the improved minimum path algorithm with backtracking, accumulated energy of a pixel point by using a second preset formula, where the second preset formula is as follows:
P_cur'＝P_cur-P_bk；
the P _ cur' represents the final accumulated energy of the pixel point, the P _ cur represents the initial accumulated energy of the pixel point, and the P _ bk represents the accumulated energy of the pixel point reached after the set step number is traced from the pixel point.
The implementation process of the functions and actions of each unit in the above device is specifically described in the implementation process of the corresponding step in the above method, and is not described herein again.
For the device embodiments, since they substantially correspond to the method embodiments, reference may be made to the partial description of the method embodiments for relevant points. The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of the application. One of ordinary skill in the art can understand and implement it without inventive effort.
The above description is only exemplary of the present application and should not be taken as limiting the present application, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present application should be included in the scope of protection of the present application.
Claims (10)
1. A method of extracting a blood vessel, the method comprising:
determining an aorta region in a medical image to be detected based on a designated pixel point, wherein the designated pixel point is positioned on an aorta;
extracting a central axis of the aorta region;
determining an end point of the central axis;
and taking the end point as a starting point, and searching from the starting point by adopting an improved minimum path algorithm with backtracking to obtain a blood vessel region in the medical image to be detected, wherein in the improved minimum path algorithm with backtracking, an energy value of a pixel point is calculated based on a saliency characteristic value of the pixel point, and the saliency characteristic value of the pixel point is positively correlated with a gray value of the pixel point.
2. The method according to claim 1, wherein determining the aorta region in the medical image to be examined based on a specified pixel point comprises:
taking a designated pixel point as an initial seed point;
and taking the initial seed point as a starting point, and performing region growth from the starting point by adopting a region growth algorithm to obtain an aorta region in the medical image to be detected.
3. The method of claim 1, wherein said determining an end point of said central axis comprises:
calculating the number of connected pixels on the central axis;
and determining the pixel points with the connection number of 1 as the end points of the central axis.
4. The method of claim 1, wherein the saliency characteristic of the pixel is calculated by:
calculating the saliency characteristic value of the pixel point by adopting a first preset formula, wherein the first preset formula is as follows:
M＝f(p_cur)*c(θ)；
wherein, M represents the saliency characteristic value of the pixel point, c (theta) represents the saliency of the pixel point in the theta direction, p _ cur represents the gray value of the pixel point, and f (p _ cur) represents a monotonically increasing nonlinear function related to the gray value of the pixel point.
5. The method according to claim 1, wherein in the improved minimum path with backtracking algorithm, a second predetermined formula is used to calculate the accumulated energy of the pixel points, and the second predetermined formula is as follows:
P_cur'＝P_cur-P_bk；
the P _ cur' represents the final accumulated energy of the pixel point, the P _ cur represents the initial accumulated energy of the pixel point, and the P _ bk represents the accumulated energy of the pixel point reached after the set step number is traced from the pixel point.
6. A blood vessel extraction device, the device comprising:
the aorta determining module is used for determining an aorta region in the medical image to be detected based on a specified pixel point, and the specified pixel point is positioned on the aorta;
the central axis extraction module is used for extracting a central axis of the aorta area;
an endpoint determination module for determining an endpoint of the central axis;
and the searching module is used for searching from the starting point by adopting an improved minimum path algorithm with backtracking to obtain a blood vessel region in the medical image to be detected, wherein in the improved minimum path algorithm with backtracking, the energy value of the pixel point is calculated based on the saliency characteristic value of the pixel point, and the saliency characteristic value of the pixel point is positively correlated with the gray value of the pixel point.
7. The apparatus of claim 6, wherein the aorta determination module comprises:
the seed point determining submodule is used for taking a specified pixel point as an initial seed point;
and the region growing sub-module is used for taking the initial seed point as a starting point and performing region growing from the starting point by adopting a region growing algorithm to obtain the aorta region in the medical image to be detected.
8. The apparatus of claim 6, wherein the endpoint determination module comprises:
the connection number calculation submodule is used for calculating the connection number of the pixel points on the central axis;
and the determining submodule is used for determining the pixel points with the connection number of 1 as the end points of the central axis.
9. The apparatus of claim 6, wherein the search module comprises:
the energy value calculating operator module is used for calculating the saliency characteristic value of the pixel point by adopting a first preset formula, and the first preset formula is as follows:
M＝f(p_cur)*c(θ)；
wherein, M represents the saliency characteristic value of the pixel point, c (theta) represents the saliency of the pixel point in the theta direction, p _ cur represents the gray value of the pixel point, and f (p _ cur) represents a monotonically increasing nonlinear function related to the gray value of the pixel point.
10. The apparatus of claim 6, wherein the search module comprises:
an accumulated energy calculation submodule for, in the improved minimum path with backtracking algorithm,
calculating the accumulated energy of the pixel points by adopting a second preset formula, wherein the second preset formula is as follows:
P_cur'＝P_cur-P_bk；
the P _ cur' represents the final accumulated energy of the pixel point, the P _ cur represents the initial accumulated energy of the pixel point, and the P _ bk represents the accumulated energy of the pixel point reached after the set step number is traced from the pixel point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810153950.6A CN109035194B (en) | 2018-02-22 | 2018-02-22 | Blood vessel extraction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810153950.6A CN109035194B (en) | 2018-02-22 | 2018-02-22 | Blood vessel extraction method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109035194A CN109035194A (en) | 2018-12-18 |
| CN109035194B true CN109035194B (en) | 2021-07-30 |
Family
ID=64629558
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810153950.6A Active CN109035194B (en) | 2018-02-22 | 2018-02-22 | Blood vessel extraction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109035194B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110288572B (en) * | 2019-06-13 | 2021-09-21 | 北京理工大学 | Automatic extraction method and device for blood vessel center line |
| CN110338830B (en) * | 2019-07-30 | 2022-12-13 | 赛诺威盛科技(北京)股份有限公司 | Method for automatically extracting head and neck blood vessel central path in CTA image |
| CN110706272B (en) * | 2019-10-17 | 2022-03-29 | 上海微创卜算子医疗科技有限公司 | Blood vessel centerline extraction method and device, electronic equipment and storage medium |
| CN110738652B (en) * | 2019-10-17 | 2023-08-11 | 青岛海信医疗设备股份有限公司 | Pulmonary artery and vein separation method and device |
| CN110796653B (en) * | 2019-10-31 | 2022-08-30 | 北京市商汤科技开发有限公司 | Image processing and neural network training method, device, equipment and medium |
| CN111429438B (en) * | 2020-03-30 | 2023-10-13 | 中国科学院深圳先进技术研究院 | Method, apparatus and storage medium for extracting blood vessel center line and heart coronary tree |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2005069223A2 (en) * | 2004-01-15 | 2005-07-28 | Algotec Systems Ltd. | Vessel centerline determination |
| CN106296664A (en) * | 2016-07-30 | 2017-01-04 | 上海联影医疗科技有限公司 | Vessel extraction method |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7912266B2 (en) * | 2007-08-07 | 2011-03-22 | Siemens Medical Solutions Usa, Inc. | System and method for robust segmentation of tubular structures in 2D and 3D images |
| US8073227B2 (en) * | 2008-05-09 | 2011-12-06 | Siemens Aktiengesellschaft | System and method for geometric modeling of tubular structures |
-
2018
- 2018-02-22 CN CN201810153950.6A patent/CN109035194B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2005069223A2 (en) * | 2004-01-15 | 2005-07-28 | Algotec Systems Ltd. | Vessel centerline determination |
| CN106296664A (en) * | 2016-07-30 | 2017-01-04 | 上海联影医疗科技有限公司 | Vessel extraction method |
Non-Patent Citations (3)
| Title |
|---|
| Centerline constrained minimal path propagation for vessel extraction;Yang Chen等;《2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI)》;20140731;全文 * |
| Curve-Like Structure Extraction Using Minimal Path Propagation With Backtracking;Yang Chen等;《IEEE Transactions on Image Processing》;20151102;全文 * |
| 基于最小路径的血管图像分割算法;陈鉴;《中国优秀硕士学位论文全文数据库医药卫生科技辑》;20170315;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109035194A (en) | 2018-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109035194B (en) | Blood vessel extraction method and device | |
| RU2413995C2 (en) | Method of improving image post processing using deformable grids | |
| US10210612B2 (en) | Method and system for machine learning based estimation of anisotropic vessel orientation tensor | |
| US10716457B2 (en) | Method and system for calculating resected tissue volume from 2D/2.5D intraoperative image data | |
| US8675943B2 (en) | Method and system for heart isolation in cardiac computed tomography volumes for patients with coronary artery bypasses | |
| US9119559B2 (en) | Method and system of generating a 3D visualization from 2D images | |
| US9014449B2 (en) | Method and system for segmentation and removal of pulmonary arteries, veins, left atrial appendage | |
| CN107067409A (en) | A kind of blood vessel separation method and system | |
| CN105096310B (en) | Divide the method and system of liver in magnetic resonance image using multi-channel feature | |
| Su et al. | Junction detection for linear structures based on Hessian, correlation and shape information | |
| DE102007018763B9 (en) | Method for arterial-venous image separation in blood pool contrast media | |
| CN108805871A (en) | Blood-vessel image processing method, device, computer equipment and storage medium | |
| US20110002520A1 (en) | Method and System for Automatic Contrast Phase Classification | |
| CN108022251A (en) | A kind of extracting method and system of the center line of tubular structure | |
| CN111768411A (en) | Coronary artery central line extraction method and device, computer equipment and storage medium | |
| JP4564233B2 (en) | Image processing apparatus, image data processing method, and program | |
| US20220036561A1 (en) | Method for image segmentation, method for training image segmentation model | |
| CN109712695A (en) | The processing method and interactive display method of blood-vessel image, computer installation | |
| CN112862835A (en) | Coronary vessel segmentation method, device, equipment and computer readable storage medium | |
| CN112102275B (en) | Pulmonary aortic vessel image extraction method, device, storage medium and electronic equipment | |
| JP2009045286A (en) | Medical image processor, medical image diagnostic apparatus, and program | |
| Singh et al. | Automatic blastomere detection in day 1 to day 2 human embryo images using partitioned graphs and ellipsoids | |
| Wieclawek et al. | Watershed based intelligent scissors | |
| JP7187680B2 (en) | LINE STRUCTURE EXTRACTION APPARATUS AND METHOD, PROGRAM AND TRAINED MODEL | |
| CN113592764A (en) | System and method for identifying and segmenting aneurysms of different anatomical regions of different modalities |
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 |