Automatic method and device for locating thrombi in the portal vein system

Abstract

The application discloses a portal vein system thrombus automatic positioning method and device, relates to the technical field of image processing, and comprises the following steps: determining the first intersection point of SMV branches, SV branches and MPV branches, and the second intersection point of MPV branches, LPV branches and RPV branches by using the skeleton line extracted from a portal vein system blood vessel three-dimensional image; determining a projection plane by using the intersection point of a circle with the first intersection point or the second intersection point as the center and a set value as the radius; projecting the three-dimensional image to the projection plane, and performing region growth calculation based on the projected image with the first intersection point or the second intersection point as the seed point until the blood vessels in the projection image are divided into three connected regions; and determining the blood vessel branch where the thrombus is located by using the pre-labeled thrombus label and the positional relationship of the three connected regions. The application can improve the accuracy and efficiency of portal vein system thrombus automatic positioning and identification.

Description

Technical Field

[0001] This invention relates to the field of medical image processing technology, and in particular to an automatic method and device for locating thrombi in the portal vein system. Background Technology

[0002] This section is intended to provide background or context for the embodiments of the invention set forth in the claims. The description herein is not an admission that it is prior art simply because it is included in this section.

[0003] In the portal venous system, accurately identifying the branch vessels affected by a thrombus is crucial. Currently, imaging techniques such as computed tomography (CT) and magnetic resonance imaging (MRI) have become the main methods for detecting thrombi in the portal venous system. However, existing technologies still have limitations in the thrombus branch localization stage. Clinically, the localization of thrombi in the portal venous system mainly relies on manual methods by physicians. Physicians need to compare the vessel course layer by layer using two-dimensional tomographic images and manually mark the portal venous branch where the thrombus is located (such as branches of the superior mesenteric vein and splenic vein). This process is time-consuming, labor-intensive, and easily influenced by subjective experience. Existing image processing algorithms based on threshold segmentation or region growing are difficult to effectively handle the tortuous course and branch angle changes of the portal venous system, while deep learning-based detection models perform poorly in the problem of branch thrombus localization. In summary, current methods for thrombus localization in the portal venous system suffer from large computer recognition errors and low efficiency, still requiring physicians to manually identify the portal venous branch where the thrombus is located. Summary of the Invention

[0004] This invention provides an automatic thrombus localization method for the portal vein system, which improves the accuracy and efficiency of automatic thrombus localization and identification in the portal vein system, eliminating the need for manual identification by doctors and improving the user experience. The method includes:

[0005] The skeleton line is extracted using a three-dimensional image of the portal vein system vessels; the coordinates of the thrombus are marked on the three-dimensional image.

[0006] The 26-neighborhood algorithm was used to calculate all endpoints and intersections of the skeleton line.

[0007] Based on the physical location relationships of multiple portal venous system vascular branches, all endpoints and intersections of the skeletal line, the vascular branches and their intersections are determined; the vascular branch intersections include the first intersection of the SMV (Superior Mesenteric Vein) branch, the SV (Splenic Vein) branch, and the MPV (Main Portal Vein) branch, and the second intersection of the MPV branch, the LPV (Left Portal Vein) branch, and the RPV (Right Portal Vein) branch;

[0008] The projection plane is determined by the intersection of a circle with the first or second intersection point as the center and a set value as the radius with the skeleton line.

[0009] The three-dimensional image is projected onto the projection plane to obtain the projected image;

[0010] Region growing calculations are performed using the first or second intersection point in the projected image as seed points until the projected image is divided into three connected regions.

[0011] By using the projected coordinates of the thrombus location in the projected image and the positional relationship of the three connected regions, the vascular branch where the thrombus is located can be determined.

[0012] This invention also provides an automatic portal vein system thrombosis localization device to improve the accuracy and efficiency of automatic thrombosis localization and identification in the portal vein system, and to improve the user experience for doctors. The device includes:

[0013] The skeleton extraction module is used to extract skeleton lines from a three-dimensional image of the portal vein system vessels; the three-dimensional image is marked with the coordinates of the thrombus.

[0014] The vascular branch bifurcation point determination module is used to calculate all endpoints and intersections of the skeleton line using a 26-neighborhood algorithm; based on the physical positional relationship of multiple portal vein system vascular branches, all endpoints and intersections of the skeleton line, it determines the vascular branches and their intersections; the vascular branch intersections include the first intersection of the SMV branch, SV branch and MPV branch, and the second intersection of the MPV branch, LPV branch and RPV branch.

[0015] The projection processing module is used to determine the projection plane by using the intersection of a circle with a set radius and a first or second intersection point as the center and the skeleton line; and to project the three-dimensional image onto the projection plane to obtain the projected image.

[0016] The blood vessel branching module is used to perform region growth calculations using the first or second intersection point in the projection image as seed points until the projection image is divided into three connected regions.

[0017] The thrombus localization module is used to determine the vascular branch where the thrombus is located by using the projected coordinates of the thrombus in the projected image and the positional relationship of the three connected regions.

[0018] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described method for automatic localization of portal vein system thrombosis.

[0019] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for automatic localization of portal vein system thrombosis.

[0020] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method for automatic localization of portal vein system thrombosis.

[0021] In this embodiment of the invention, vascular branches and their intersections are determined based on the portal vein system vascular skeleton line, including the first intersection of SMV, SV, and MPV branches, and the second intersection of MPV, LPV, and RPV branches. Then, an optimal projection plane is determined by intersecting the skeleton line with a circle centered at the first or second intersection point and having a set radius. The three-dimensional image is then projected onto the projection plane to obtain a projected image. This transforms the complex three-dimensional spatial branch structure into a two-dimensional projected image. Based on the two-dimensional projected image, vascular branches are divided and matched with the pre-marked coordinates of the thrombus location, achieving precise localization of the specific portal vein branch where the thrombus is located. This solves the problems of high subjectivity and low efficiency caused by relying on manual interpretation by doctors in existing technologies, improving the doctor's user experience. At the same time, the implementation method is relatively simple and has higher localization processing efficiency compared to methods such as deep learning. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0023] Figure 1This is a flowchart illustrating the automatic thrombus localization method for the portal vein system in an embodiment of the present invention;

[0024] Figure 2 This is a specific example diagram of the automatic thrombus localization method for the portal vein system in an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of the vascular structure and skeleton of the portal vein system in an embodiment of the present invention;

[0026] Figure 4 This is a schematic diagram of key points for dividing blood vessel branches in an embodiment of the present invention;

[0027] Figure 5 This is a schematic diagram of thrombus localization in an embodiment of the present invention;

[0028] Figure 6 This is a schematic diagram of an automatic thrombus localization device for the portal vein system in an embodiment of the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0030] The acquisition, storage, use, and processing of data in this application all comply with the relevant provisions of national laws and regulations.

[0031] To facilitate a clear description of the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish the same or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order.

[0032] To address the issues of low efficiency and insufficient accuracy in branch localization of portal vein system thrombosis in existing technologies, this invention proposes a projection-based automatic thrombosis localization method for the portal vein system. By constructing a three-dimensional projection model of the portal vein system vessels, the complex spatial branching structure is transformed into a two-dimensional projection, enabling precise localization of the specific branch of the portal vein where the thrombus is located. This method achieves automatic localization of portal vein system thrombosis, solving the problems of high subjectivity and low efficiency in manual interpretation in existing technologies, and providing efficient and reliable technical support for the precise diagnosis and treatment of portal vein system thrombosis.

[0033] Figure 1 This is a flowchart illustrating the automatic thrombus localization method for the portal vein system in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes:

[0034] Step 101: Extract the skeleton line using a three-dimensional image of the portal vein system vessels; the three-dimensional image is marked with the coordinates of the thrombus.

[0035] Step 102: Calculate all endpoints and intersections of the skeleton line using the 26-neighborhood algorithm;

[0036] Step 103: Based on the physical location relationship of multiple portal vein system vascular branches, all endpoints and intersections of the skeleton line, determine the vascular branches and the vascular branch intersections; the vascular branch intersections include the first intersection of the SMV branch, SV branch and MPV branch, and the second intersection of the MPV branch, LPV branch and RPV branch.

[0037] Step 104: Determine the projection plane by using the intersection of a circle with the first or second intersection point as the center and a set value as the radius with the skeleton line;

[0038] Step 105: Project the 3D image onto the projection plane to obtain the projected image;

[0039] Step 106: Perform region growing calculations using the first or second intersection point in the projected image as seed points until the projected image is divided into three connected regions.

[0040] Step 107: Using the projected coordinates of the thrombus location in the projected image and the positional relationship of the three connected regions, determine the vascular branch where the thrombus is located.

[0041] The following is about Figure 1 The method for automatic thrombus localization in the portal vein system is explained in detail.

[0042] The main idea of ​​this invention is based on the projection technology from three-dimensional space to two-dimensional space. The portal vein system vessels are projected onto the optimal projection plane. The key points for dividing the vessel branches are determined on the projection plane. The SMV branches, SV branches, MPV branches, LPV branches, and RPV branches of the vessels are obtained. The thrombus is matched with each branch region, and finally the automatic localization of the thrombus is achieved.

[0043] Figure 2 This is a specific example diagram of the automatic thrombus localization method for the portal vein system in an embodiment of the present invention, as shown below. Figure 2 As shown, the method includes:

[0044] Step 201: Obtain labeled images of blood vessels and thrombi;

[0045] Step 202: Locate the bifurcation points of blood vessels;

[0046] Step 203: Determine the projection plane and perform projection;

[0047] Step 204: Determine the key points for dividing blood vessel branches;

[0048] Step 205: Thrombus localization.

[0049] During the procedure, three-dimensional images of the portal vein system vessels are first acquired, such as abdominal CECT (Abdominal Contrast-Enhanced Computed Tomography) images of the patient.

[0050] In this embodiment of the invention, blood vessels and thrombi are labeled based on three-dimensional images of the portal venous system. For example, adaptive threshold segmentation is used to identify blood vessels and thrombi by utilizing the density difference between blood vessels and surrounding tissues. For instance, neural networks such as U-Net (Convolutional Networks for Biomedical Image Segmentation) and 3D U-Net (Learning Dense Volumetric Segmentations from Sparse Annotations) are used to automatically label blood vessels and thrombi. Alternatively, morphological operations and connected component analysis are combined to label blood vessels and thrombi.

[0051] By marking blood vessels and thrombi, the coordinates of the thrombus can be determined.

[0052] In step 101, the skeleton line is extracted using a three-dimensional image of the portal vein system.

[0053] For example, an iterative erosion operation can be performed on an image using a predefined structuring element template to remove boundary points until only the middle line remains.

[0054] In practice, the skeleton lines of the portal vein system vessels can be extracted using the skeletonize function in the skimage library (an open-source library in Python specifically for image processing).

[0055] Next, the branching points of blood vessels are found. In step 102, all endpoints and intersections of the skeleton line are calculated using the 26-neighborhood algorithm. During implementation, the first intersection of the SMV branch, SV branch and MPV branch is searched first.

[0056] In one embodiment, calculating all endpoints and intersections of the skeleton line using a 26-neighborhood algorithm may include:

[0057] If there is only one neighboring pixel within the 26-neighborhood of the current pixel, then the current pixel is determined as the endpoint.

[0058] If there are 3 or more neighboring pixels within the 26-neighborhood of the current pixel, the current pixel is determined to be an intersection point;

[0059] Remove small branches of the skeleton lines; the small branches are defined as those with a pixel count of less than or equal to 3 pixels from the intersection point to the endpoint.

[0060] In this process, removing small branches of the skeleton line can be achieved by directly assigning the pixel value at the small branch to 0, thus preventing it from interfering with the search for blood vessel bifurcation points.

[0061] In step 103, the vascular branches and their intersections are determined based on the physical location relationships of multiple portal venous system vascular branches, all endpoints and intersections of the skeleton line.

[0062] During implementation, the actual physical location relationships of SMV branches, SV branches, MPV branches, LPV branches, and RPV branches are matched with all endpoints and intersections to determine the vascular branches on the skeleton line and the intersections of the vascular branches.

[0063] In one embodiment, determining the vascular branches and their intersections based on the physical location relationships of multiple portal venous system vascular branches, all endpoints and intersections of the skeletal line, may include:

[0064] Starting from the endpoint with the smallest x and z coordinates, traverse along the skeleton line to find the first intersection point as the first intersection point of the SMV branch, SV branch and MPV branch; where the x-axis direction is defined to increase from left to right, the z-axis direction to increase from bottom to top, and the y-axis direction is perpendicular to the xz plane, that is, perpendicular to the screen viewed by the user.

[0065] Starting from the first intersection point, traverse along the non-SMV and non-SV branches on the skeleton line to find the first intersection point as the second intersection point of the MPV, LPV, and RPV branches.

[0066] Figure 3 This is a schematic diagram of the portal vein system vascular structure and skeleton lines in an embodiment of the present invention, for reference only. Figure 3 First, all endpoints and intersections of the skeleton line are calculated using a 26-neighborhood. Endpoints are found when a pixel has only one neighbor within its 26-neighborhood, and intersections are found when a pixel has at least three neighbors within its 26-neighborhood. Then, small branches of the skeleton line are removed. Finally, starting from the lower left endpoint (the endpoint with the smallest x and z coordinates), the nearest intersection is found by traversing the skeleton line; this intersection is the first point of intersection between the SMV, SV, and MPV branches. Figure 3As shown by the red dot, the second intersection of the MPV branch, LPV branch, and RPV branch is then found. This process is similar to finding the first intersection of the SMV branch, SV branch, and MPV branch. The difference is that the starting point is the first intersection of the SMV branch, SV branch, and MPV branch.

[0067] Next, the projection plane is determined. In step 104, the projection plane is determined by the intersection of a circle with the first or second intersection point as the center and a set value as the radius with the skeleton line.

[0068] During implementation, the projection planes of the SMV branch, SV branch, and MPV branch need to be calculated separately from the projection planes of the MPV branch, LPV branch, and RPV branch.

[0069] The calculation process for the projection plane involves using the first or second intersection point as the center and setting an empirical value as the search radius, typically between 5 and 10. Within this radius, three key points intersecting with the skeleton line are further identified. A plane is then determined according to formula (1), which is the projection plane.

[0070] Ax + By + Cz + D = 0 (1)

[0071] In formula (1), A, B, C, and D are the parameters of the projection plane.

[0072] In step 105, the 3D image is projected onto the projection plane to obtain the projected image.

[0073] In one embodiment, to facilitate subsequent calculations and visualization, and to make it easier for users to view, the three-dimensional image is projected onto a projection plane to obtain a projected image. This may include: determining a rotation matrix from the projection plane to the coordinate plane; rotating the three-dimensional image according to the rotation matrix; and projecting the rotated three-dimensional image onto the coordinate plane to obtain a projected image.

[0074] In one embodiment, the coordinate plane is the xz plane, the rotation matrix R is represented by the following formula (2), and the projection process is shown in formula (3).

[0075] (2)

[0076] F(x,y,z)=(x,z) (3)

[0077] in, It is the identity matrix. Here, is the rotation angle, n is the normal vector of the projection plane (i.e., n = (A, B, C), and t is the normal vector of the xz plane (i.e., t = (0, 1, 0)). It is the unit vector of the rotation axis.

[0078] Formula (3) F(x,y,z)=(x,z) means that the y coordinate value is removed during projection.

[0079] Then, the key points for dividing blood vessel branches are determined. In step 106, region growing calculations are performed using the first or second intersection point in the projected image as seed points until the projected image is divided into three connected regions.

[0080] During implementation, based on the region self-growth algorithm, the bifurcation point is used as the seed point until the blood vessels in the projected image are successfully divided into three connected regions.

[0081] In step 107, the vascular branch where the thrombus is located is determined by using the projected coordinates of the thrombus location in the projected image and the positional relationship of the three connected regions.

[0082] Figure 4 This is a schematic diagram of key points for dividing blood vessel branches in an embodiment of the present invention, such as... Figure 4 As shown, the red dot is used as the seed point to divide three connected regions. Then, the findContours function in the OpenCV library is used to accurately extract the contour lines from the selected regions. Subsequently, the minimum distance from the red dot to each point on all contour lines is calculated, and the three closest points are selected as the key points for branching.

[0083] In one embodiment, determining the vascular branch where the thrombus is located by using the projected coordinates of the thrombus in the projected image and the positional relationship of the three connected regions may include:

[0084] Based on the three connected regions, the blood vessel branches in the projected image are divided into a first part, a second part, and a third part; wherein, when the projection plane is determined using the first intersection point, the first part is an SMV branch, the second part is an SV branch, and the third part is a combination of other branches except for the SMV and SV branches; when the projection plane is determined using the second intersection point, the first part is an LPV branch, the second part is an RPV branch, and the third part is a combination of other branches except for the LPV and RPV branches.

[0085] The coordinates of the thrombus are matched with the first part, the second part, and the third part respectively to determine the vascular branch where the thrombus is located; wherein, when the vascular branch where the thrombus is located is the third part, the projection plane is determined again by the first or second intersection point that was not used when determining the current projection plane, and the vascular branch where the thrombus is located is determined by the newly determined projection plane.

[0086] Figure 5 This is a schematic diagram of thrombus localization in an embodiment of the present invention, as shown below. Figure 5 As shown, taking SMV branch, SV branch, and MPV branch as examples, based on Figure 4The three key points are to divide the projected blood vessels in 2D space into SMV branches, SV branches, and combinations of other branches. During the branching process, after removing the triangular region formed by the three points, the centroid coordinates of the remaining three connected regions are used to determine whether it belongs to an SMV branch, an SV branch, or a combination of MPV, RPV, and LPV. After obtaining the precise branching, the projection coordinates of the thrombus determine which branch coordinate region it falls within, thus identifying which branch the thrombus belongs to.

[0087] When the thrombus belongs to a combination of branches other than the SMV branch and SV branch, a second projection is required. The first projection uses the intersection of the first intersection point as the center and the skeleton line as the radius to determine the projection plane. The second projection uses the intersection of the second intersection point as the center and the skeleton line as the radius to determine the projection plane. The remaining specific steps are the same as steps 104 to 107.

[0088] In summary, the embodiments of the present invention map the portal vein system vessels from 3D space to 2D space through projection, and divide the vessel branches by finding key points for branch division in 2D space, thereby achieving a precise, efficient, and objective automatic thrombus localization function in the portal vein system. Experimental verification shows that the accuracy rate can reach 90%.

[0089] This invention also provides an automatic portal vein system thrombosis localization device, as described in the following embodiments. Since the principle by which this device solves the problem is similar to the automatic portal vein system thrombosis localization method, the implementation of this device can refer to the implementation of the automatic portal vein system thrombosis localization method; repeated details will not be elaborated further.

[0090] Figure 6 This is a schematic diagram of the automatic portal vein system thrombosis localization device in an embodiment of the present invention, as shown below. Figure 6 As shown, the device 600 includes:

[0091] The skeleton extraction module 601 is used to extract skeleton lines from a three-dimensional image of the portal vein system vessels; the three-dimensional image is marked with the coordinates of the thrombus.

[0092] The vascular branch bifurcation point determination module 602 is used to calculate all endpoints and intersections of the skeleton line using a 26-neighborhood algorithm; and to determine the vascular branches and their intersections based on the physical positional relationships of multiple portal vein system vascular branches and all endpoints and intersections of the skeleton line; the vascular branch intersections include the first intersection of the SMV branch, SV branch and MPV branch, and the second intersection of the MPV branch, LPV branch and RPV branch.

[0093] The projection processing module 603 is used to determine the projection plane by using the intersection of a circle with a set value as the center and a set value as the radius with the skeleton line; and to project the three-dimensional image onto the projection plane to obtain the projected image.

[0094] The blood vessel branching module 604 is used to perform region growth calculations using the first or second intersection point in the projection image as seed points until the projection image is divided into three connected regions.

[0095] The thrombus localization module 605 is used to determine the vascular branch where the thrombus is located by using the projected coordinates of the thrombus in the projected image and the positional relationship of the three connected regions.

[0096] In one embodiment, the blood vessel branch bifurcation point determination module 602 is specifically used for:

[0097] If there is only one neighboring pixel within the 26-neighborhood of the current pixel, then the current pixel is determined as the endpoint.

[0098] If there are 3 or more neighboring pixels within the 26-neighborhood of the current pixel, the current pixel is determined to be an intersection point;

[0099] Remove small branches of the skeleton lines; the small branches are defined as those with a pixel count of less than or equal to 3 pixels from the intersection point to the endpoint.

[0100] In one embodiment, the blood vessel branch bifurcation point determination module 602 is specifically used for:

[0101] Starting from the endpoint with the smallest x and z coordinates, traverse along the skeleton line to find the first intersection point as the first intersection point of the SMV branch, SV branch and MPV branch; where the x-axis direction is defined to increase from left to right, the z-axis direction to increase from bottom to top, and the y-axis direction is perpendicular to the xz plane.

[0102] Starting from the first intersection point, traverse along the non-SMV and non-SV branches on the skeleton line to find the first intersection point as the second intersection point of the MPV, LPV, and RPV branches.

[0103] In one embodiment, the projection processing module 603 is specifically used for:

[0104] Determine the rotation matrix from the projection plane to the coordinate plane;

[0105] Rotate the 3D image according to the rotation matrix;

[0106] The rotated 3D image is projected onto the coordinate plane to obtain the projected image.

[0107] In one embodiment, the coordinate plane is the xz plane, and the rotation matrix is ​​represented as follows:

[0108] ;

[0109] Where R is the rotation matrix, It is the identity matrix. Where is the rotation angle, n is the normal vector of the projection plane, and t is the normal vector of the xz plane. It is the unit vector of the rotation axis.

[0110] In one embodiment, the thrombus localization module 605 is specifically used for:

[0111] Based on the three connected regions, the blood vessel branches in the projected image are divided into a first part, a second part, and a third part; wherein, when the projection plane is determined using the first intersection point, the first part is an SMV branch, the second part is an SV branch, and the third part is a combination of other branches except for the SMV and SV branches; when the projection plane is determined using the second intersection point, the first part is an LPV branch, the second part is an RPV branch, and the third part is a combination of other branches except for the LPV and RPV branches.

[0112] The coordinates of the thrombus are matched with the first part, the second part, and the third part respectively to determine the vascular branch where the thrombus is located; wherein, when the vascular branch where the thrombus is located is the third part, the projection plane is determined again by the first or second intersection point that was not used when determining the current projection plane, and the vascular branch where the thrombus is located is determined by the newly determined projection plane.

[0113] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described method for automatic localization of portal vein system thrombosis.

[0114] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for automatic localization of portal vein system thrombosis.

[0115] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method for automatic localization of portal vein system thrombosis.

[0116] In this embodiment of the invention, vascular branches and their intersections are determined based on the portal vein system vascular skeleton line, including the first intersection of SMV, SV, and MPV branches, and the second intersection of MPV, LPV, and RPV branches. Then, using the first or second intersection point as the center and a circle with a set radius as the intersection point with the skeleton line, an optimal projection plane is determined. The three-dimensional image is projected onto the projection plane to obtain a projected image. This transforms the complex three-dimensional spatial branch structure into a two-dimensional projected image. Based on the two-dimensional projected image, vascular branches are divided and matched with the pre-marked coordinates of the thrombus, achieving precise localization of the specific portal vein branch where the thrombus is located. This solves the problems of high subjectivity and low efficiency in existing technologies that rely on manual interpretation by doctors, improving the doctor's user experience. At the same time, the implementation method is relatively simple and has higher localization processing efficiency compared to methods such as deep learning.

[0117] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0118] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0119] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0120] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0121] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for automatic positioning of a thrombus in a portal vein system, characterized in that, The method comprises the steps of: extracting a skeleton line by using a three-dimensional image of a portal vein system blood vessel; the three-dimensional image is marked with a thrombus coordinate; calculating all end points and intersection points of the skeleton line by using a 26-neighborhood algorithm; determining a blood vessel branch and a blood vessel branch intersection point according to a physical position relationship of a plurality of portal vein system blood vessel branches, all end points and intersection points of the skeleton line; the blood vessel branch intersection point comprises a first intersection point of a superior mesenteric vein SMV branch, a splenic vein SV branch and a main portal vein MPV branch, and a second intersection point of the MPV branch, a left portal vein LPV branch and a right portal vein RPV branch; determining a projection plane by using a circular intersection point of a circle with the first intersection point or the second intersection point as a center and a set value as a radius and the skeleton line; projecting the three-dimensional image onto the projection plane to obtain a projection image; performing region growing calculation on the projection image by taking the first intersection point or the second intersection point as a seed point until the projection image is divided into three connected regions; determining a blood vessel branch where the thrombus is located by using a projection coordinate of the thrombus coordinate in the projection image and a position relationship of the three connected regions; wherein the determination of the blood vessel branch where the thrombus is located by using the projection coordinate of the thrombus coordinate in the projection image and the position relationship of the three connected regions comprises: based on the three connected regions, the blood vessel branch in the projection image is divided into a first part, a second part and a third part; when the first intersection point is used to determine the projection plane, the first part is the SMV branch, the second part is the SV branch, and the third part is a combination of other branches except the SMV branch and the SV branch; when the second intersection point is used to determine the projection plane, the first part is the LPV branch, the second part is the RPV branch, and the third part is a combination of other branches except the LPV branch and the RPV branch; matching the thrombus coordinate with the first part, the second part and the third part respectively to determine the blood vessel branch where the thrombus is located; when the blood vessel branch where the thrombus is located is the third part, the first intersection point or the second intersection point which is not used in determining the current projection plane is used again to determine the projection plane, and the blood vessel branch where the thrombus is located is determined by using the projection plane determined again.

2. The method of claim 1, wherein, The calculation of all end points and intersection points of the skeleton line by using the 26-neighborhood algorithm comprises: when there is only one neighbor pixel in the 26-neighborhood of the current pixel, the current pixel is determined as an end point; when there are more than or equal to three neighbor pixels in the 26-neighborhood of the current pixel, the current pixel is determined as an intersection point; removing a small branch of the skeleton line; the small branch is a branch with a pixel number less than or equal to three pixels from an intersection point to an end point.

3. The method of claim 1, wherein, The determination of the blood vessel branch and the blood vessel branch intersection point according to the physical position relationship of the plurality of portal vein system blood vessel branches, all end points and intersection points of the skeleton line comprises: taking an end point with the smallest x coordinate and z coordinate as a starting point, and finding a first intersection point along the skeleton line as the first intersection point of the SMV branch, the SV branch and the MPV branch; wherein the x axis direction is defined as increasing from left to right, the z axis direction is defined as increasing from bottom to top, and the y axis direction is perpendicular to the xz plane. Taking the first intersection point as a starting point, a first intersection point is found along a direction of a non-SMV branch, a non-SV branch on the skeleton line to be a second intersection point of the MPV branch, the LPV branch and the RPV branch.

4. The method of claim 3, wherein, The three-dimensional image is projected onto a projection plane to obtain a projection image, comprising: determining a rotation matrix of the projection plane to the coordinate plane; rotating the three-dimensional image according to the rotation matrix; projecting the rotated three-dimensional image onto the coordinate plane to obtain the projection image.

5. The method of claim 4, wherein, The coordinate plane is an xz plane, and the rotation matrix is represented as follows: ； where R is a rotation matrix, is an identity matrix, is a rotation angle, n is a projection plane normal vector, t is an xz plane normal vector, is a unit vector of a rotation axis.

6. A device for automatic positioning of a thrombus in a portal vein system, characterized in that comprising: The skeleton extraction module is configured to extract a skeleton line by using a three-dimensional image of a portal vein system blood vessel, and the three-dimensional image is labeled with a blood clot coordinate; The blood vessel branch bifurcation point determination module is configured to calculate all end points and intersection points of the skeleton line by using a 26-neighborhood algorithm, and determine blood vessel branches and blood vessel branch intersection points according to a physical position relationship of the multiple portal vein system blood vessel branches, the all end points and the intersection points of the skeleton line, wherein the blood vessel branch intersection points comprise a first intersection point of an SMV branch, an SV branch and an MPV branch, and a second intersection point of the MPV branch, an LPV branch and an RPV branch; The projection processing module is configured to determine a projection plane by using an intersection point of a circle with a radius of a set value and the skeleton line, taking the first intersection point or the second intersection point as a center of the circle, project the three-dimensional image onto the projection plane to obtain a projection image, and determine a blood clot coordinate in the projection image. The blood vessel branch division module is configured to perform region growing calculation by taking the first intersection point or the second intersection point in the projection image as a seed point until the projection image is divided into three connected regions. The blood clot positioning module is configured to determine a blood vessel branch where the blood clot is located by using a position relationship between the blood clot coordinate in the projection image and the three connected regions. The blood clot positioning module is configured to: based on the three connected regions, divide the blood vessel branch in the projection image into a first part, a second part and a third part; when the first intersection point is used to determine the projection plane, the first part is the SMV branch, the second part is the SV branch, and the third part is a combination of other branches except the SMV branch and the SV branch; when the second intersection point is used to determine the projection plane, the first part is the LPV branch, the second part is the RPV branch, and the third part is a combination of other branches except the LPV branch and the RPV branch; match the blood clot coordinate with the first part, the second part and the third part respectively to determine the blood vessel branch where the blood clot is located; when the blood vessel branch where the blood clot is located is the third part, the projection plane is determined again by using the first intersection or the second intersection point that is not used when the current projection plane is determined, and the blood vessel branch where the blood clot is located is determined by using the projection plane determined again.

7. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the method of any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the method of any one of claims 1 to 5.

9. A computer program product, characterised in that, The computer program product comprises a computer program, and the computer program is executed by the processor to implement the method of any one of claims 1 to 5.

Images