Medical image processing method, device, medium, and product

Abstract

This invention discloses a medical image processing method, device, medium, and product, relating to the field of medical image processing technology. The method includes: acquiring the segmentation results of hemangioma and tumor-bearing vessels in a scanned image of a target vessel; determining a first point set for the hemangioma based on the hemangioma segmentation results; determining a second point set for the tumor-bearing vessels based on the tumor-bearing vessel segmentation results; determining a bidirectional neighboring point set for the first and second point sets, each bidirectional neighboring point set including at least three non-collinear pixels; determining a plane equation for the tumor diameter plane based on the bidirectional neighboring point set; determining the intersection points of the plane containing the plane equation with the first and second point sets, and using the plane enclosed by all intersection points as the tumor diameter plane. The technical solution provided by this invention solves the technical problem of slow speed in existing tumor diameter plane determination methods, improving the speed of tumor diameter plane determination.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of medical image processing technology, and in particular to a medical image processing method, device, medium and product. Background Technology

[0002] Hemangioma refers to an irreversible and continuous expansion of blood vessels, resulting in cystic or spindle-shaped tumor-like dilatation of the vessel wall.

[0003] Studies have shown that the risk of hemangioma rupture is significantly correlated with its morphological parameters, family history, and biomechanical indicators. Morphological parameters can be quickly calculated by combining patient impact data and geometric modeling techniques, and are currently an important means commonly used in clinical practice to predict the risk of hemangioma rupture.

[0004] Determining the neck plane is a crucial step in calculating the relevant morphological parameters of intracranial hemangiomas. Currently, morphological assessment in clinical settings is mainly based on two-dimensional computed tomography (CTA) and magnetic resonance imaging (MRA) techniques. The neck plane is obtained through manual measurement or semi-automatic measurement using relevant software, and the relevant morphological parameters are calculated. However, this method is relatively slow.

[0005] Therefore, improving the speed of determining the tumor diameter plane has become an urgent technical problem to be solved. Summary of the Invention

[0006] This invention provides a medical image processing method, device, medium, and product to improve the speed of determining the tumor diameter plane.

[0007] According to one aspect of the present invention, a medical image processing method is provided, the method comprising: Obtain the segmentation results of hemangioma and tumor-bearing vessels from the blood vessel scanning image of the target object, and determine a first point set for the hemangioma based on the segmentation results of the hemangioma, and determine a second point set for the tumor-bearing vessels based on the segmentation results of the tumor-bearing vessels. Determine a bidirectional neighbor set for the first point set and the second point set, wherein the bidirectional neighbor set includes at least three non-collinear pixels; The plane equation for the tumor diameter plane is determined based on the bidirectional nearest neighbor set; Determine the intersection points of the plane containing the plane equation with the first point set and the second point set, and take the plane enclosed by all the intersection points as the tumor diameter plane.

[0008] According to another aspect of the present invention, a medical image processing apparatus is provided, the apparatus comprising: The point set determination module is used to acquire the hemangioma segmentation result and the tumor-bearing vessel segmentation result of the target object blood vessel scanning image, and determine a first point set for the hemangioma based on the hemangioma segmentation result, and determine a second point set for the tumor-bearing vessel based on the tumor-bearing vessel segmentation result. The neighbor set module is used to determine a bidirectional neighbor set for the first point set and the second point set, wherein the bidirectional neighbor set includes at least three non-collinear pixels; A plane equation module is used to determine a plane equation for the tumor diameter plane based on the bidirectional neighbor point set; The tumor diameter plane module is used to determine the intersection points of the plane containing the plane equation with the first point set and the second point set, and to take the plane enclosed by all the intersection points as the tumor diameter plane.

[0009] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: One or more processors; Storage device for storing one or more programs. When one or more programs are executed by one or more processors, the one or more processors implement any of the medical image processing methods described in the embodiments of the present invention.

[0010] According to another aspect of the present invention, a computer-readable storage medium is provided, which stores computer instructions for causing a processor to execute and implement any of the medical image processing methods of the present invention.

[0011] According to another aspect of the present invention, a computer program product is provided, which, when executed by a processor, implements any of the medical image processing methods described in the embodiments of the present invention.

[0012] The technical solution of this invention involves image segmentation of a target object's blood vessel scanning image to determine the segmentation results of hemangioma and the tumor-bearing vessel. Then, based on the hemangioma segmentation results, a first set of points for the hemangioma is determined, and based on the tumor-bearing vessel segmentation results, a second set of points for the tumor-bearing vessel is determined, along with a bidirectional neighbor set of points for the first and second set. Finally, based on this bidirectional neighbor set, a plane equation for the tumor diameter plane is determined, and the region enclosed by the intersections of the plane containing this equation with the first and second point sets is taken as the tumor diameter plane. This achieves the technical efficiency of automatically determining the tumor diameter plane. Furthermore, the determination of the tumor diameter plane is not limited by the size and type of hemangioma, effectively overcoming the technical limitations of traditional manual delineation and measurement methods, such as strong subjectivity, poor repeatability, and low efficiency. Moreover, since the bidirectional neighbor set includes at least three non-collinear pixels, and the number of pixels included is much smaller than the number of pixels included in the union of the first and second point sets, and each pixel in the bidirectional neighbor set is distributed in the tumor diameter plane or its neighborhood, the plane equation of the tumor diameter plane can be quickly determined based on this bidirectional neighbor set, thereby quickly determining the tumor diameter plane.

[0013] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0015] Figure 1 A schematic flowchart of a medical image processing method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of hemangioma segmentation results provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the tumor-bearing vessel segmentation results provided in an embodiment of the present invention; Figure 4 A schematic diagram of the intersection points of the first point set, the second point set, and the target plane provided in an embodiment of the present invention; Figure 5 This is a schematic diagram illustrating the positional relationship between the first neighbor set, the second neighbor set, and the target plane, provided in an embodiment of the present invention. Figure 6 A schematic diagram illustrating the relationship between the edge line of the tumor diameter plane and the plane containing the plane equation, provided in an embodiment of the present invention. Figure 7 This is a schematic diagram of the tumor diameter provided in an embodiment of the present invention; Figure 8 This is another schematic flowchart of the medical image processing method provided in an embodiment of the present invention; Figure 9 A schematic flowchart of the image segmentation method provided in an embodiment of the present invention; Figure 10A This is a schematic diagram of the target blood vessel segmentation result provided in an embodiment of the present invention; Figure 10B This is a schematic diagram of the tumor-bearing vascular mesh and the hemangioma mesh provided in an embodiment of the present invention; Figure 11A This is a schematic diagram of the structure of a medical image processing device provided in an embodiment of the present invention; Figure 11B This is another schematic diagram of the medical image processing device provided in an embodiment of the present invention; Figure 12 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0016] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0017] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0018] Figure 1This is a flowchart illustrating a medical image processing method provided in an embodiment of the present invention. This embodiment is applicable to the automatic determination of tumor diameter planes. The method can be executed by an image processing device, which can be implemented in hardware and / or software and can be configured in electronic devices such as computers or servers. Figure 1 As shown, the method in this embodiment includes: S110. Obtain the segmentation results of hemangioma and tumor-bearing vessels in the blood vessel scanning image of the target object, and determine a first point set for the hemangioma based on the segmentation results of the hemangioma, and determine a second point set for the tumor-bearing vessels based on the segmentation results of the tumor-bearing vessels.

[0019] Vascular scan images refer to medical images obtained by scanning a predetermined area of ​​a target object using imaging techniques specifically designed for blood vessels. Examples include two-dimensional computed tomography (CTA) and magnetic resonance imaging (MR). The predetermined area can be the head or other parts of the body.

[0020] In one embodiment, a blood vessel scan image is segmented based on an existing image segmentation method to obtain a hemangioma segmentation result (see [link]). Figure 2 ) and the segmentation results of the tumor-bearing vessels (see Figure 3 The segmentation results for hemangiomas include only the hemangioma itself, excluding the blood vessel. The segmentation results for tumor-bearing vessels include only the tumor-bearing vessel, excluding the hemangioma and other blood vessels on it. A tumor-bearing vessel is a blood vessel on which a hemangioma grows. For example, if a hemangioma L grows on blood vessel T of target object A, then blood vessel T is the tumor-bearing vessel of hemangioma L.

[0021] The pixel sets belonging to hemangiomas in the hemangioma segmentation results are used as the first point set, and the pixel sets belonging to tumor-bearing vessels in the tumor-bearing vessel segmentation results are used as the second point set.

[0022] S120. Determine a bidirectional neighboring point set for the first point set and the second point set, wherein the bidirectional neighboring point set includes at least three non-collinear pixels.

[0023] The nearest neighbor set refers to finding the "closest" point in space for one or more points in a given dataset.

[0024] The bidirectional neighbor set includes a first neighbor set in the first point set relative to the second point set, and a second neighbor set in the second point set relative to the first point set; each pixel in the bidirectional neighbor set is determined based on the same distance threshold, and the number of pixels in the bidirectional neighbor set is greater than or equal to 3.

[0025] It is understandable that the larger the distance threshold, the more pixels there are in the bidirectional neighbor set.

[0026] In one embodiment, the bidirectional neighbor set is determined by the following steps: Step a1: Determine the first neighboring point set in the first point set relative to the second point set, and the second neighboring point set in the second point set relative to the first point set, based on the first distance threshold.

[0027] The first distance threshold is an initial threshold, which is usually small, such as 0.1 mm. It is determined whether there are pixels in the first point set whose distance to the second point set is less than or equal to the first distance threshold. If so, these pixels are added to the first neighboring point set. Similarly, it is determined whether there are pixels in the second point set whose distance to the first point set is less than or equal to the second threshold. If so, these pixels are added to the second neighboring point set. The union of the first and second neighboring point sets is determined, and the number of pixels in this union is determined to be greater than or equal to a predetermined threshold. If so, this union is used as a bidirectional neighboring point set; otherwise, step a2 is executed.

[0028] Step a2: If the number of pixels in the union of the first neighbor set and the second neighbor set is less than a predetermined threshold, then update the first distance threshold based on a predetermined step size. The predetermined threshold is greater than or equal to 3, and the updated first distance threshold is greater than the original first distance threshold.

[0029] If the union of pixels is less than a predetermined threshold, the first distance threshold is updated based on a predetermined step size. For example, if the predetermined step size is 0.05mm, the current first distance threshold + 0.05mm is used as the updated first distance threshold. This increases the number of pixels in the union by increasing the first distance threshold.

[0030] Step a3: Return to the step of determining the first neighbor set and the second neighbor set based on the first distance threshold, until the number of pixels in the union of the first neighbor set and the second neighbor set is greater than or equal to a predetermined threshold.

[0031] After the first distance threshold is updated, based on the updated first distance threshold, return to step a3 until the number of pixels in the union is greater than or equal to the predetermined threshold.

[0032] For this embodiment, exemplarily: the geometric point set of the hemangioma is recorded as... There are a total of m points, and the point set on the tumor-bearing vessel is recorded as follows: There are n points in total. Introducing the minimum distance formula, we calculate... and The minimum distance from each pixel in the dataset to each other is used to obtain the first and second neighbor sets. An initial distance threshold is set to 0.1mm. or If the distance to a pixel is less than a certain threshold, then the pixel's ID and spatial coordinates are saved. In other words, the number of pixels in its corresponding first neighbor set is recorded as follows: ,for In other words, the number of pixels in its corresponding second neighbor set is recorded as follows: .when If the distance threshold is increased at 0.05mm intervals, it continues until the number of pixels in the union of the first and second nearest neighbor sets reaches 10. The minimum distance formula can be the Euclidean distance formula, the Manhattan distance formula, etc.

[0033] S130. Determine the plane equation for the tumor diameter plane based on the bidirectional neighbor point set.

[0034] In one embodiment, a bidirectional neighbor set is input into a pre-trained machine learning model to obtain a plane equation for the tumor diameter plane.

[0035] Among them, the normal vector of the plane equation is preferably a unit normal vector.

[0036] S140. Determine the intersection points of the plane containing the plane equation with the first point set and the second point set, and take the plane enclosed by all the intersection points as the apex plane.

[0037] After determining the plane equation, determine the intersection points of the target plane with the first point set and the second point set (see...). Figure 4 The plane region enclosed by all intersection points is the tumor diameter plane, where the target plane is the plane containing the plane equation.

[0038] Figure 4 Specifically, it shows the first set of points corresponding to the hemangioma, the second set of points corresponding to the tumor-bearing vessels, the intersection of the plane containing the plane equation with the first set of points, and the intersection of the plane containing the plane equation with the second set of points.

[0039] from Figure 5 It can be seen that the number of intersections between the first set of points and the target plane (the plane containing the plane equation) is far greater than the number of intersections in the first set of neighboring points that are not located on the target plane. These intersections are located in the adjacent... Figure 5 The first target pixel (pixels not located on the target plane in the first neighboring point set) appears as the projection point on the target plane; the number of intersections between the second point set and the target plane is much greater than the number of intersections in the second neighboring point set that are not located on the target plane. These intersections are located in the adjacent... Figure 5 The projection point of the second target pixel (a pixel in the second neighboring point set that is not distributed on the target plane) onto the target plane appears. This further illustrates the accuracy of the target plane determined by the embodiments of the present invention, and the accuracy of the target plane can significantly improve the accuracy of the determined tumor diameter plane.

[0040] from Figure 5 It can also be seen that the intersections of the first point set with the target plane and the intersections of the second point set with the target plane form an irregular circular curve. Fitting this irregular circular curve yields... Figure 6 The midline is the outer edge of the plane of the tumor diameter.

[0041] The outer edge and its internal area constitute the tumor diameter plane. (See details...) Figure 7 The black area in the text.

[0042] The technical solution provided in this invention segmentes a blood vessel scan image of a target object to determine the segmentation results of the hemangioma and the tumor-bearing vessel. Then, based on the hemangioma segmentation results, a first set of points for the hemangioma is determined, and based on the tumor-bearing vessel segmentation results, a second set of points for the tumor-bearing vessel is determined, along with a bidirectional neighbor set of points for the first and second set of points. Finally, based on this bidirectional neighbor set, a plane equation for the tumor diameter plane is determined, and the region enclosed by the intersections of the plane containing this equation with the first and second set of points is taken as the tumor diameter plane. This achieves the automatic determination of the tumor diameter plane. The method is effective, and the determination of the tumor diameter plane is not limited by the size and type of hemangioma, effectively overcoming the technical limitations of traditional manual delineation and measurement methods, such as strong subjectivity, poor repeatability, and low efficiency. Moreover, since the bidirectional neighbor set includes at least three non-collinear pixels, and the number of pixels included is much smaller than the number of pixels included in the union of the first and second point sets, and each pixel in the bidirectional neighbor set is distributed in the tumor diameter plane or its neighborhood, the plane equation of the tumor diameter plane can be quickly determined based on this bidirectional neighbor set, thereby quickly determining the tumor diameter plane.

[0043] Based on the aforementioned embodiments, after the tumor diameter plane is determined, the morphological parameters of the hemangioma are determined according to the tumor diameter plane, the first point set, and the second point set.

[0044] These morphological parameters include, but are not limited to, the long diameter of the hemangioma, the first transverse diameter perpendicular to the long diameter, the height, the second transverse diameter perpendicular to the height, the diameter of the hemangioma, the longest diameter, the length-to-width ratio, the height-to-width ratio, the inflow angle, and the non-spherical index.

[0045] Specifically, the center of the hemangioma plane and the Euclidean distance from each pixel in the first point set to the center are determined, and the maximum Euclidean distance is taken as the major axis of the hemangioma; the plane passing through each pixel of the major axis and perpendicular to the major axis is determined, and the maximum diameter of all planes is taken as the first transverse axis of the hemangioma; the second intersection point of the straight line passing through each pixel of the hemangioma plane and along the normal vector direction with the first point set is determined, and the maximum distance between all second intersection points and the hemangioma plane is taken as the height of the hemangioma; the height perpendicular vector is determined, and the height perpendicular vector that has two intersection points with the first point set is determined, and the maximum intersection value of all height perpendicular vectors with the hemangioma is determined, and the maximum intersection value is taken as the second transverse axis of the hemangioma; the longest axis of the hemangioma plane is taken as the diameter of the hemangioma; the distance between each pair of pixels on the hemangioma wall is determined, and the maximum distance is taken as the longest diameter of the hemangioma.

[0046] The ratio of the aforementioned length to the first transverse diameter is taken as the length-to-width ratio of the hemangioma; the ratio of the aforementioned height to the second transverse diameter is taken as the height-to-width ratio of the hemangioma; the blood flow direction and diameter of the carrier vessel are determined based on the segmentation results of the carrier vessel; the inflow angle of the hemangioma is determined based on the blood flow direction and the aforementioned long diameter; the ratio of the aforementioned long diameter to the flow diameter is taken as the length-to-short ratio; the surface area and volume of the hemangioma are calculated, and the surface area of ​​a sphere with the same volume as the hemangioma is determined; the difference between the surface area of ​​the hemangioma and the surface area of ​​the sphere with the same volume as the hemangioma is determined, and the ratio of this difference to the surface area of ​​the sphere is taken as the non-spherical index of the hemangioma.

[0047] Figure 8 This is another schematic flowchart of the medical image processing method provided in this embodiment of the invention. This embodiment is used to refine the process of determining the plane equation in the foregoing embodiments. The technical solution of this embodiment can be combined with other embodiments; for the same or related parts, they can be described in conjunction with the descriptions of other embodiments, and will not be repeated here. Figure 8 As shown, the method in this embodiment may specifically include: S210. Obtain the segmentation results of hemangioma and tumor-bearing vessels in the blood vessel scanning image of the target object, and determine the first point set for the hemangioma based on the segmentation results of the hemangioma, and determine the second point set for the tumor-bearing vessels based on the segmentation results of the tumor-bearing vessels.

[0048] S220. Determine a bidirectional neighbor set for the first point set and the second point set, wherein the bidirectional neighbor set includes at least three non-collinear pixels.

[0049] S2301. Determine the first normal vector of the apex plane based on the geometric properties of the bidirectional neighboring point set.

[0050] The geometric properties of a point set are used to describe the shape, structure, and distribution characteristics of the point set in space.

[0051] In one embodiment, the centroid of the bidirectional neighbor set and the covariance matrix corresponding to the centroid are determined; principal component analysis is performed on the covariance matrix to obtain three orthogonal eigenvectors and their eigenvalues; the first normal vector of the tumor diameter plane is determined based on the eigenvalues ​​of the three orthogonal eigenvectors.

[0052] Specifically, principal component analysis is used to find the main distribution directions of the covariance matrix by analyzing the variance in different directions. Eigenvectors define the main directions of the distribution, and eigenvalues ​​represent the distribution intensity in these directions. In this embodiment, the direction with the smallest eigenvalue, i.e., the direction where the point set is "thinnest," is preferably used as the first normal vector.

[0053] Of course, the first normal vector of the tumor diameter plane corresponding to the bidirectional neighboring point set can also be determined by methods such as least squares fitting and random sampling consistency.

[0054] S2302. Under the condition that the constraint equation holds, by minimizing the sum of the distances between each pixel in the bidirectional neighbor set and the plane containing the general form plane equation, the second normal vector and intercept for the general form plane equation are determined, the constraint equation is, and the dot product of the first normal vector and the second normal vector is a predetermined value.

[0055] First, define the initial general form of the plane equation: ; in, The second normal vector, Let be the intercept. Next, minimize the sum of distances from all pixels in the bidirectional neighbor set to the plane containing the general plane equation, i.e., find . The minimum value of , where, This represents the number of pixels included in the bidirectional neighbor set. To be identified as The distance from a pixel to the plane containing the general formula for the plane equation.

[0056] Optimization variables include and the distance variable corresponding to each pixel in the bidirectional neighbor set. ,common One variable.

[0057] Establish constraints for each pixel. The inequality constraints on the distance to the plane containing the general equation of the plane are as follows: ; To ensure that the optimized normal vector is consistent with the initial estimate, a constraint equation is added, specifically: ; in, The first normal vector, It is a predetermined value, preferably 1, to improve the calculation speed, but it can also be other values.

[0058] S2303. By normalizing the second normal vector and the intercept, the second normal vector becomes a unit vector, and the plane equation of the plane containing the tumor diameter plane is obtained.

[0059] The normalized result of the second normal vector is used as the target normal vector; the ratio of the intercept to the magnitude of the second normal vector is used as the target intercept.

[0060] Specifically, ; ; in, For the target normal vector, The target intercept.

[0061] Once the target normal vector and target intercept are determined, they are substituted into the general plane equation to obtain a plane equation, which is the plane equation of the plane containing the tumor diameter plane.

[0062] S240. Determine the intersection points of the plane containing the plane equation with the first point set and the second point set, and take the plane enclosed by all the intersection points as the apex plane.

[0063] The technical solution of this invention determines the first normal vector of the tumor diameter plane based on the geometric properties of the bidirectional neighbor set. A constraint equation is constructed to constrain the first and second normal vectors. Under the condition that the constraint equation holds, the second normal vector and intercept for the general plane equation are determined by minimizing the sum of distances from each pixel in the bidirectional neighbor set to the plane containing the general plane equation. This improves the accuracy of determining the second normal vector and the accuracy of the tumor diameter plane determined based on the second normal vector. By normalizing the second normal vector and intercept, the second normal vector becomes a unit vector, thus obtaining the plane equation of the plane containing the tumor diameter plane. This reduces the computational load of determining the tumor diameter plane based on the plane equation and improves the speed of determining the tumor diameter plane.

[0064] Figure 9 This is a schematic flowchart of the image segmentation method provided in an embodiment of the present invention. This embodiment is used to refine the process of determining the segmentation results of hemangiomas and the segmentation results of the tumor-bearing vessels in the foregoing embodiments. The technical solution of this embodiment can be combined with other embodiments. For the same or related parts, they can be described in conjunction with the descriptions of other embodiments, and will not be repeated here. Figure 9 As shown, the method in this embodiment may specifically include: S3101. Obtain a vascular scan image of the target object, including the hemangioma and the tumor-bearing vessel.

[0065] Taking two-dimensional computed tomography (CTA) imaging as an example, prospective ECG-gated scanning can be performed according to a preset scanning protocol (65 ml of contrast agent is injected at a rate of 5.0 ml / s, and 30 ml of normal saline is injected at the same rate). During the scan, adaptive sequence prospective ECG-gated triggering technology is activated. If the patient's heart rate is ≤70 beats / min, the full-dose window is set at 65%-75% of the RR interval; if the patient's heart rate is >70 beats / min, the full-dose window is set at 35%-45% of the RR interval.

[0066] S3102. Input the blood vessel scan image into the pre-trained first segmentation model to obtain the target blood vessel segmentation result, which includes the tumor-bearing blood vessel and the hemangioma growing on the tumor-bearing blood vessel.

[0067] This embodiment uses a network model to segment blood vessel scan images, specifically: First, the thin-slice DICOM format vascular scan images are preprocessed. Preprocessing mainly includes addressing inconsistencies in scanning parameters, smoothing noise, and normalizing data signals, such as "data resampling," "data signal intensity normalization," and "image denoising." This effectively normalizes the image, removes noise, and maintains clear boundaries, laying the groundwork for subsequent image segmentation.

[0068] Next, radiologists with over 5 years of clinical experience manually segmented the preprocessed CTA images to obtain the segmentation results. The segmentation results include the segmentation of the target vessel and the segmentation of the hemangioma.

[0069] The target blood vessel segmentation result is used as the first label of the blood vessel scan image to generate the first training sample, and the hemangioma segmentation result is used as the second label of the blood vessel scan image to generate the second training sample. Following the above process, the first training sample for each blood vessel scan image in the combination of blood vessel scan images, and the second training sample for each target blood vessel segmentation result in the combination of target blood vessel segmentation results, are determined, thus obtaining the first training sample combination and the second training sample combination.

[0070] The segmentation model is trained using the first combination of training samples to obtain a pre-trained first segmentation model; the segmentation model is trained using the second combination of training samples to obtain a pre-trained second segmentation model. After the pre-trained first segmentation model is determined, the blood vessel scan image is input into the pre-trained first segmentation model to obtain the target blood vessel segmentation result.

[0071] The segmentation model is a 3DUnet neural network, which includes non-local blocks to incorporate relevant information about the characteristics around blood vessels into the calculation, thereby expanding the sensing area for feature acquisition.

[0072] The blood vessel scan image is input into the pre-trained first segmentation model to obtain the target blood vessel segmentation result (see...). Figure 10A ).

[0073] S3103. Input the target blood vessel segmentation result into the pre-trained second segmentation model to obtain the hemangioma segmentation result.

[0074] The target blood vessel segmentation result is input into the pre-trained second segmentation model so that the second segmentation model can extract the hemangioma from the target blood vessel segmentation result and obtain the hemangioma segmentation result.

[0075] By combining the pre-trained first segmentation model with the pre-trained second segmentation model, the hemangioma was segmented step by step, which improved the accuracy of hemangioma segmentation compared to segmenting hemangiomas directly from blood vessel scan images.

[0076] S3104. Determine the segmentation result of the tumor-bearing vessel based on the segmentation results of the target vessel and the segmentation result of the hemangioma.

[0077] In one embodiment, after the target vessel segmentation result and the hemangioma segmentation result are determined, the difference between the target vessel segmentation result and the hemangioma segmentation result is determined based on Boolean operation, and the difference is used as the segmentation result of the tumor-bearing vessel.

[0078] In another embodiment, the complete vascular mesh and the hemangioma sub-mesh are determined, a unified vertex index is established, and an adjacency graph is constructed. The average edge length of the mesh is calculated for adaptive threshold setting. A K-dimensional tree (KDTree) is used to quickly match the corresponding vertices of the hemangioma sub-mesh with those of the complete mesh. A dynamic threshold is set (default 0.9 × average edge length), and all vertices within the distance threshold are marked as hemangioma regions (UIA). Fault tolerance mechanism: if the threshold is too strict and no matching vertices are found, the threshold is automatically lowered to the 10th percentile of the distance distribution as the new threshold.

[0079] Adjacency analysis is used to locate the boundary vertices between the hemangioma and the carrier vessel (i.e., vertices adjacent to UIA vertices but not marked as UIA). Based on the user-defined proximal and distal bandwidths, non-UIA vertices within the distance range are extracted to form candidate carrier vessel regions. All vessel branches connected to the boundary seed point are preserved.

[0080] Starting from the boundary seed point, Dijkstra's algorithm is used to calculate the shortest geodesic distance to all vertices. The principal direction is analyzed using PCA, and only the branch with the highest consistency with the seed point direction is retained (the included angle threshold is adjustable). If the initial bandwidth does not extract a valid region, the bandwidth is gradually increased (2 × 8 × average side length) until successful.

[0081] The extracted tumor-bearing vessel vertices are smoothed using Laplacian smoothing, with an adjustable number of iterations (default 10) to eliminate jagged edges. Triangular faces are selected based on the vertex mask, retaining faces containing at least two tumor-bearing vessel vertices to ensure mesh integrity. Tumor-bearing vessels are then generated (STL / PLY / OBJ format). See [link to documentation]. Figure 10B The darker parts in the text.

[0082] In this embodiment of the invention, the use of a pre-trained first segmentation model in conjunction with a pre-trained second segmentation model achieves accurate segmentation of the target blood vessel and the hemangioma, thereby improving the accuracy of the segmentation result of the tumor-bearing blood vessel determined based on the segmentation results of the target blood vessel and the hemangioma.

[0083] Figure 11A This is a schematic diagram of a medical image processing device provided in an embodiment of the present invention. The medical image processing device can be used in electronic devices. Figure 11A As shown, the medical image processing device includes: The point set determination module 410 is used to acquire the hemangioma segmentation result and the tumor-bearing vessel segmentation result of the target object blood vessel scanning image, and determine a first point set for the hemangioma based on the hemangioma segmentation result, and determine a second point set for the tumor-bearing vessel based on the tumor-bearing vessel segmentation result. The neighbor point set module 420 is used to determine a bidirectional neighbor point set for the first point set and the second point set, wherein the bidirectional neighbor point set includes at least three non-collinear pixels. Plane equation module 430 is used to determine the plane equation for the tumor diameter plane based on the bidirectional neighbor point set; The tumor diameter plane module 440 is used to determine the intersection points of the plane containing the plane equation with the first point set and the second point set, and to take the plane enclosed by all the intersection points as the tumor diameter plane.

[0084] The technical solution of the image processing device provided in this embodiment of the invention involves image segmentation of a target object's blood vessel scanning image to determine the segmentation results of hemangioma and the tumor-bearing vessel. Then, based on the hemangioma segmentation results, a first point set for the hemangioma is determined, and based on the tumor-bearing vessel segmentation results, a second point set for the tumor-bearing vessel is determined, along with a bidirectional neighboring point set for the first and second point sets. Finally, based on this bidirectional neighboring point set, a plane equation for the tumor diameter plane is determined, and the region enclosed by the intersections of the plane containing this plane equation with the first and second point sets is taken as the tumor diameter plane. This achieves automatic determination of the tumor diameter plane. The technology achieves a superior effect, and the determination of the tumor diameter plane is not limited by the size and type of hemangioma, effectively overcoming the limitations of traditional manual delineation and measurement methods, which are characterized by strong subjectivity, poor repeatability, and low efficiency. Moreover, since the bidirectional neighbor set includes at least three non-collinear pixels, and the number of pixels included is much smaller than the number of pixels included in the union of the first and second point sets, and each pixel in the bidirectional neighbor set is distributed in the tumor diameter plane or its neighborhood, the plane equation of the tumor diameter plane can be quickly determined based on this bidirectional neighbor set, thereby quickly determining the tumor diameter plane.

[0085] In one embodiment, the plane equation module 430 includes: The first normal vector unit is used to determine the first normal vector of the tumor diameter plane based on the geometric properties of the bidirectional neighbor point set. The second normal vector unit is used to determine the second normal vector and intercept for the general form plane equation by minimizing the sum of distances between each pixel in the bidirectional neighbor set and the plane containing the general form plane equation, under the condition that the constraint equation holds. The constraint equation is that the dot product of the first normal vector and the second normal vector is a predetermined value. The plane equation unit is used to obtain the plane equation of the plane containing the tumor diameter plane by normalizing the second normal vector and the intercept to make the second normal vector a unit vector.

[0086] In one embodiment, the first normal vector unit is used for: Determine the centroid of the bidirectional neighbor set and the covariance matrix corresponding to the centroid; Principal component analysis was performed on the covariance matrix to obtain three orthogonal eigenvectors and their eigenvalues. The first normal vector of the tumor diameter plane is determined based on the eigenvalues ​​of the three orthogonal eigenvectors.

[0087] In one embodiment, the bidirectional neighbor set includes a first neighbor set of the first set of points relative to the second set of points, and a second neighbor set of the second set of points relative to the first set of points; Each pixel in the bidirectional neighbor set is determined based on the same distance threshold, and the number of pixels in the bidirectional neighbor set is greater than or equal to 3.

[0088] In one embodiment, the neighbor point set module 420 is used for: Based on a first distance threshold, a first neighboring set of points for the second set of points in the first set of points and a second neighboring set of points for the first set of points in the second set of points are determined. If the number of pixels in the union of the first neighbor set and the second neighbor set is less than a predetermined threshold, the first distance threshold is updated based on a predetermined step size. The predetermined threshold is greater than or equal to 3, and the updated first distance threshold is greater than the original first distance threshold. The process returns to the step of determining a first neighboring set of points for the second set of points in the first set of points, and a second neighboring set of points for the first set of points in the second set of points, based on a first distance threshold, until the number of pixels in the union of the first neighboring set and the second neighboring set of points is greater than or equal to the predetermined threshold.

[0089] In one embodiment, such as Figure 11B As shown, the device also includes a parameter module 450, which is used for: The morphological parameters of the hemangioma are determined based on the tumor diameter plane, the first point set, and the second point set.

[0090] In one embodiment, the point set determination module 410 is used for: Acquire a vascular scan image of the target object, the vascular scan image including the hemangioma and the tumor-bearing vessel; The blood vessel scan image is input into a pre-trained first segmentation model to obtain the target blood vessel segmentation result, which includes the tumor-bearing vessel and the hemangioma growing on the tumor-bearing vessel. The target blood vessel segmentation result is input into a pre-trained second segmentation model to obtain the hemangioma segmentation result; The segmentation result of the tumor-bearing vessel is determined based on the segmentation result of the target vessel and the segmentation result of the hemangioma.

[0091] The medical image processing apparatus provided in the embodiments of the present invention can execute the medical image processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0092] It is worth noting that the various units and modules included in the above-mentioned medical image processing device are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the protection scope of the embodiments of the present invention.

[0093] Figure 12 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the embodiments of the invention described and / or claimed herein.

[0094] like Figure 12 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0095] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0096] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as medical image processing methods.

[0097] In some embodiments, the medical image processing method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or mounted on electronic device 10 via read-only memory (ROM) 12 and / or communication unit 19. When the computer program is loaded into random access memory (RAM) 13 and executed by processor 11, one or more steps of the medical image processing method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the medical image processing method by any other suitable means (e.g., by means of firmware).

[0098] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0099] Computer programs for implementing the medical image processing methods of embodiments of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The computer programs can be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0100] This invention provides a computer-readable storage medium storing computer instructions for causing a processor to execute a medical image processing method, including: Obtain the segmentation results of hemangioma and tumor-bearing vessels from the blood vessel scanning image of the target object, and determine a first point set for the hemangioma based on the segmentation results of the hemangioma, and determine a second point set for the tumor-bearing vessels based on the segmentation results of the tumor-bearing vessels. Determine a bidirectional neighbor set for the first point set and the second point set, wherein the bidirectional neighbor set includes at least three non-collinear pixels; The plane equation for the tumor diameter plane is determined based on the bidirectional nearest neighbor set; Determine the intersection points of the plane containing the plane equation with the first point set and the second point set, and use the plane enclosed by the intersection points as the tumor diameter plane.

[0101] In the context of embodiments of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0102] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0103] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0104] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0105] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from ROM 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of the embodiments of the present invention.

[0106] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements a medical image processing method according to any embodiment of the invention.

[0107] In implementing a computer program product, computer program code for performing the operations of embodiments of the present invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0108] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the embodiments of the present invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of the embodiments of the present invention can be achieved, and this document does not impose any restrictions.

[0109] The specific embodiments described above do not constitute a limitation on the scope of protection of the embodiments of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the embodiments of the present invention should be included within the scope of protection of the embodiments of the present invention.

Claims

1. A medical image processing method, characterized in that, Obtain the segmentation results of hemangioma and tumor-bearing vessels from the blood vessel scanning image of the target object, and determine a first point set for the hemangioma based on the segmentation results of the hemangioma, and determine a second point set for the tumor-bearing vessels based on the segmentation results of the tumor-bearing vessels. Determine a bidirectional neighbor set for the first point set and the second point set, wherein the bidirectional neighbor set includes at least three non-collinear pixels; The plane equation for the tumor diameter plane is determined based on the bidirectional nearest neighbor set; Determine the intersection points of the plane containing the plane equation with the first point set and the second point set, and take the plane enclosed by all the intersection points as the tumor diameter plane.

2. The method according to claim 1, characterized in that, The determination of the plane equation for the tumor diameter plane based on the bidirectional neighbor set includes: The first normal vector of the tumor diameter plane is determined based on the geometric properties of the bidirectional neighbor point set; Under the condition that the constraint equation holds, the second normal vector and intercept for the general form plane equation are determined by minimizing the sum of the distances between each pixel in the bidirectional neighbor set and the plane containing the general form plane equation. The constraint equation is that the dot product of the first normal vector and the second normal vector is a predetermined value. By normalizing the second normal vector and the intercept, making the second normal vector a unit vector, the plane equation of the plane containing the tumor diameter plane is obtained.

3. The method according to claim 2, characterized in that, Determining the first normal vector of the tumor diameter plane based on the geometric properties of the bidirectional neighbor set includes: Determine the centroid of the bidirectional neighbor set and the covariance matrix corresponding to the centroid; Principal component analysis was performed on the covariance matrix to obtain three orthogonal eigenvectors and their eigenvalues. The first normal vector of the tumor diameter plane is determined based on the eigenvalues ​​of the three orthogonal eigenvectors.

4. The method according to claim 1, characterized in that, The bidirectional neighbor set includes a first neighbor set in the first point set relative to the second point set, and a second neighbor set in the second point set relative to the first point set; Each pixel in the bidirectional neighbor set is determined based on the same distance threshold, and the number of pixels in the bidirectional neighbor set is greater than or equal to 3.

5. The method according to claim 4, characterized in that, Determining the bidirectional neighbor set of points for the first point set and the second point set includes: Based on a first distance threshold, a first neighboring set of points for the second set of points in the first set of points and a second neighboring set of points for the first set of points in the second set of points are determined. If the number of pixels in the union of the first neighbor set and the second neighbor set is less than a predetermined threshold, the first distance threshold is updated based on a predetermined step size. The predetermined threshold is greater than or equal to 3, and the updated first distance threshold is greater than the original first distance threshold. The process returns to the step of determining a first neighboring set of points for the second set of points in the first set of points, and a second neighboring set of points for the first set of points in the second set of points, based on a first distance threshold, until the number of pixels in the union of the first neighboring set and the second neighboring set of points is greater than or equal to the predetermined threshold.

6. The method according to claim 1, characterized in that, After determining the intersection point of the plane containing the planar equation and the result of the tube tumor segmentation, and using the plane enclosed by the intersection point as the tumor diameter plane, the method further includes: The morphological parameters of the hemangioma are determined based on the tumor diameter plane, the first point set, and the second point set.

7. The method according to claim 1, characterized in that, The acquisition of the hemangioma segmentation result and the tumor-bearing vessel segmentation result from the target object's blood vessel scanning image includes: Acquire a vascular scan image of the target object, the vascular scan image including the hemangioma and the tumor-bearing vessel; The blood vessel scan image is input into a pre-trained first segmentation model to obtain the target blood vessel segmentation result, which includes the tumor-bearing vessel and the hemangioma growing on the tumor-bearing vessel. The target blood vessel segmentation result is input into a pre-trained second segmentation model to obtain the hemangioma segmentation result; The segmentation result of the tumor-bearing vessel is determined based on the segmentation result of the target vessel and the segmentation result of the hemangioma.

8. An electronic device, characterized in that, The electronic device includes: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the medical image processing method as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the medical image processing method according to any one of claims 1-7.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the medical image processing method according to any one of claims 1-7.

Images