Auxiliary path planning method for inferior vena cava filter implantation puncture target

Abstract

The invention relates to the technical field of image analysis, in particular to an auxiliary path planning method for implanting an inferior vena cava filter into a puncture target. A plurality of CT images of the inferior vena cava of a patient are obtained, and a to-be-distinguished area is preliminarily divided based on vascular morphological changes. And then, in combination with shape features and area change conditions of the to-be-distinguished region, determining a blood vessel overlapping characterization evaluation index, and distinguishing an overlapping region and a non-overlapping region. Thirdly, combining the overlapping area and the non-overlapping area, analyzing shape similarity and overlapping representation evaluation index similarity, determining a smooth continuation index, and screening out a matching combination based on the index and a blood vessel overlapping area proportion; besides, gray value differences of different levels of blood vessels in the CT image are also considered, and level evaluation is performed on a non-blood vessel overlapping region, so that the region belonging to the same blood vessel is represented more accurately. And finally, more accurate puncture target navigation path planning is carried out on the region with blood vessel overlapping according to a hierarchical evaluation result, so that the planning precision is improved.

Description

Technical Field

[0001] This invention relates to the field of image analysis technology, specifically to an auxiliary path planning method for the insertion of an inferior vena cava filter into the puncture target point. Background Technology

[0002] Inferior vena cava filter placement is an effective method to prevent pulmonary embolism caused by dislodged deep vein thrombosis in the lower extremities. Accurately identifying the puncture target point is crucial during this procedure to ensure the filter is placed precisely in the intended location within the inferior vena cava. Due to the complexity of the inferior vena cava and its surrounding structures, and the varying anatomical features of different patients, path planning-assisted techniques have emerged to improve placement accuracy and safety. These techniques, combining modern imaging technologies (such as CT and ultrasound) with computer-aided navigation systems, can guide surgeons in real-time planning of the puncture target path during the procedure, avoiding damage to surrounding vital blood vessels and organs.

[0003] When planning the navigation path for the placement of a retrievable inferior vena cava filter, existing technologies typically rely on the shortest distance between puncture targets. However, in cases with complex vascular structures or overlapping vessels, simple shortest distance calculations cannot effectively avoid these complex structures, reducing the accuracy of auxiliary confirmation of the placement target of the retrievable inferior vena cava filter, easily leading to misjudgment, and affecting the accuracy of the final navigation path. Summary of the Invention

[0004] To address the technical problem that simple shortest distance calculations are ineffective in avoiding complex structures and overlapping vessels, leading to misjudgments and affecting the accuracy of navigation paths, this invention aims to provide an auxiliary path planning method for inferior vena cava filter placement puncture target points. The specific technical solution adopted is as follows:

[0005] Multiple CT images of the patient's inferior vena cava were acquired, and the vascular region was segmented in each CT image;

[0006] In each CT image, the regions to be distinguished are divided based on the change in the distance between the two edge lines of each blood vessel in the vascular region; the shape features of each region to be distinguished in each CT image and the area change between each region to be analyzed and the corresponding region in other CT images are analyzed to determine the vascular overlap characterization evaluation index of each region to be distinguished, thereby distinguishing vascular overlap regions and non-vascular overlap regions.

[0007] In each CT image, each overlapping vascular region is combined with a non-overlapping vascular region to obtain multiple region combinations. Within each region combination, the shape similarity between overlapping and non-overlapping vascular regions and the similarity between vascular overlap characterization evaluation indicators in non-overlapping vascular regions are analyzed to determine the smoothness continuation index of the region combination. The differences in the smoothness continuation index among all region combinations are analyzed, and combined with the proportion of overlapping vascular regions in the region combinations, matching combinations are selected from all region combinations.

[0008] In each CT image, the difference in gray values ​​between non-vascular overlapping regions is analyzed in the matching combination corresponding to each overlapping region of blood vessels. The non-vascular overlapping regions are then subjected to hierarchical evaluation to obtain the hierarchical evaluation results. The navigation path for the puncture target is planned based on the hierarchical evaluation results.

[0009] Furthermore, the method for obtaining the region to be distinguished includes:

[0010] In each CT image, select one of the two edge lines of each blood vessel as the target line.

[0011] On the target line, feature points are obtained based on the SURF feature point detection algorithm. A straight line perpendicular to the target line is drawn through each feature point, and the intersection of the line with another edge line is used as the reference point corresponding to each feature point.

[0012] The distance between each feature point and its corresponding reference point is used as the distance factor.

[0013] On the target line, a feature point that does not exist in the feature point set is randomly selected as the test point, and the test point corresponds to a feature point set.

[0014] Starting from the point to be measured, the feature points on both sides are traversed along the target line. For each feature point traversed, the difference in distance factor between the feature point and the feature points in the feature point set is analyzed to obtain the merging index. When the merging index meets the preset conditions, the traversed feature points are merged into the feature point set of the point to be measured, and the traversal continues; if the conditions are not met, the traversal stops.

[0015] The preset condition is set to the merged index being greater than or equal to the preset merged threshold.

[0016] On the target line, randomly select a feature point that does not exist in the feature point set as the test point and repeat the above traversal process until all feature points exist in the feature point set, and then stop to obtain the set of all feature points on the target line.

[0017] In each set of feature points, the two feature points farthest apart on the target line are taken as endpoints. The area enclosed by the line segments between the endpoints and their corresponding reference points, the line segments of the reference points on the edge line, and the line segments of the two endpoints on the target line is taken as a region to be distinguished.

[0018] Furthermore, the method for obtaining the merged indicators includes:

[0019] Each feature point encountered during the traversal is treated as a point to be analyzed.

[0020] Calculate the absolute value of the difference between the distance factor of the point to be analyzed and the distance factor of each feature point in the set of feature points corresponding to the point to be measured, and use it as the difference factor;

[0021] The mean values ​​of all differential factors corresponding to the points to be analyzed are negatively correlated and normalized, and then used as the merged index.

[0022] Furthermore, the method for obtaining the evaluation indicators for vascular overlap includes:

[0023] Feature points in different CT images are matched to determine the corresponding region of each region to be distinguished in each CT image in other CT images, and the corresponding region of each region to be distinguished in each CT image in other CT images is used as the comparison region.

[0024] In each region to be distinguished and in all corresponding comparison regions, the difference between the area of ​​the largest region and the area of ​​the smallest region is used as the first overlap characterization evaluation factor for each region to be distinguished.

[0025] In each region to be distinguished, the line connecting the midpoint of the line segment between the feature point and the corresponding reference point is taken as the center line. The mean of the curvature values ​​at all midpoints on each center line is calculated as the curvature value, and the variance of the curvature values ​​at all midpoints on each center line is taken as the fluctuation value.

[0026] The normalized value of the product of the curvature value and the fluctuation value corresponding to each region to be distinguished is used as the second overlapping characterization evaluation factor for each region to be distinguished.

[0027] The normalized sum of the first and second overlapping characterization evaluation factors for each region to be distinguished is used as the vascular overlapping characterization evaluation index for each region to be distinguished.

[0028] Furthermore, the distinction between overlapping vascular regions and non-overlapping vascular regions includes:

[0029] When the evaluation index of the vascular overlap characterization of a certain region to be distinguished is greater than the preset overlap threshold, the region to be distinguished is considered to be a vascular overlap region.

[0030] When the evaluation index of the vascular overlap characterization of a certain region to be distinguished is less than or equal to the preset overlap threshold, the region to be distinguished is considered to be a non-vascular overlap region.

[0031] Furthermore, the method for obtaining the region combination includes:

[0032] In each CT image, for any overlapping blood vessel region, the non-overlapping blood vessel region connected to the overlapping blood vessel region is selected as a candidate region.

[0033] In the overlapping region of the blood vessels, select one candidate region in each of the two directions along the edge of the blood vessel to form a region combination with the overlapping region of the blood vessel, and obtain all non-repeating region combinations.

[0034] Furthermore, the method for obtaining the smoothing continuation index includes:

[0035] In each region combination, the intersection point between the straight line containing the center line of each non-vascular overlapping region and the straight line containing the center line of the vascular overlapping region is taken as the intersection point;

[0036] The Euclidean distance between the midpoint of the maximum curvature value on the centerline of each non-vascular overlapping region and the corresponding intersection point is used as the position deviation factor. The sum of the position deviation factors corresponding to two non-vascular overlapping regions is negatively correlated and mapped to the value of the smoothing factor.

[0037] The absolute value of the difference between the vascular overlap characterization evaluation indexes of two non-vascular overlap regions is negatively correlated and normalized, and then used as the similarity factor.

[0038] The normalized sum of the smoothing factor and similarity factor for each region combination is used as the smoothing continuation index for each region combination.

[0039] Furthermore, the method for obtaining the matching combination includes:

[0040] In each CT image, for any overlapping blood vessel region, the proportion of the smooth continuation index of each region combination corresponding to the overlapping blood vessel region in the smooth continuation index of all region combinations corresponding to the overlapping blood vessel region is used as the first matching contribution.

[0041] In each region combination corresponding to the overlapping blood vessel region, the average length of the two edge lines of the blood vessels in the overlapping blood vessel region is calculated as the first length factor. The average length of the two edge lines of the blood vessels in each non-overlapping blood vessel region is calculated as the second length factor. The proportion of the first length factor in the sum of the first length factor and all second length factors is used as the second matching contribution.

[0042] The normalized sum of the first and second matching contributions for each region combination corresponding to the overlapping blood vessel region is used as the matching coefficient.

[0043] Among all the regions in the overlapping blood vessel region, the regions with a matching coefficient greater than the preset matching threshold are selected as the matching combinations.

[0044] Furthermore, the method for obtaining the hierarchical evaluation results includes:

[0045] In all matching combinations, in all non-vascular overlapping regions along each direction of the blood vessel in the overlapping region of blood vessels, the mean gray value of all pixels in each non-vascular overlapping region is calculated as the gray value feature value of each non-vascular overlapping region. All non-vascular overlapping regions are sorted in ascending order according to their gray value feature values ​​to obtain a sorted sequence.

[0046] In the two sorting sequences corresponding to the overlapping vascular regions, the two non-overlapping vascular regions with the same index value belong to the same level of blood vessels.

[0047] Furthermore, the method for obtaining the vascular region includes:

[0048] Each CT image is used as input to a pre-trained neural network, which then outputs the vascular region in each CT image.

[0049] The present invention has the following beneficial effects:

[0050] When planning the navigation path for the inferior vena cava filter placement puncture target, the accuracy of the navigation path planning may be affected by the possibility of overlapping blood vessels. Therefore, this invention mainly analyzes the impact of blood vessel overlap. First, multiple CT images of the patient's inferior vena cava are acquired, and the vascular region is segmented. Since the morphology changes when blood vessels overlap, that is, the distance between the edge lines of the blood vessels changes significantly, the vascular region is initially divided based on this feature, resulting in regions to be distinguished. Furthermore, when blood vessels overlap, the area of ​​the overlap may show different changes in different CT images. Therefore, by combining the shape characteristics of each region to be analyzed and the area changes of the region to be analyzed with the corresponding regions in other CT images, a vascular overlap characterization evaluation index is determined for the regions to be distinguished, and the overlapping and non-overlapping vascular regions are distinguished based on this index. In order to determine which non-overlapping vascular regions on both sides of the overlapping vascular region may belong to the same blood vessel, this invention combines the overlapping and non-overlapping vascular regions, resulting in multiple region combinations. Given that regions belonging to the same blood vessel should exhibit relatively consistent and smooth spatial continuity and morphology, the shape similarity between overlapping and non-overlapping vascular regions, as well as the similarity between overlap characterization evaluation indicators of non-overlapping vascular regions, were analyzed within each region combination to determine the smoothness continuity index of the region combination. Since a larger proportion of overlapping vascular regions corresponds to a smaller proportion of non-overlapping vascular regions, thus reducing matching difficulty, their weight should be increased. Therefore, based on the differences in smoothness continuity indicators and the proportion of overlapping vascular regions, matching combinations were selected across all region combinations. Regions within each matching combination represent a higher probability of belonging to the same blood vessel. Furthermore, since blood vessels at different levels exhibit different grayscale values ​​in CT images, this invention analyzes the differences in grayscale values ​​between non-overlapping vascular regions for each matching combination corresponding to an overlapping vascular region. This allows for hierarchical evaluation of the non-overlapping vascular regions, yielding hierarchical evaluation results. These results can more accurately characterize which regions belong to the same blood vessel. Consequently, for regions with overlapping vascular regions, the navigation path for the puncture target can be planned more precisely based on the hierarchical evaluation results. Attached Figure Description

[0051] To more clearly illustrate the technical solutions and advantages 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.

[0052] Figure 1This is a flowchart of an auxiliary path planning method for inferior vena cava filter placement puncture target point provided in an embodiment of the present invention;

[0053] Figure 2 This is a schematic diagram illustrating the division of a vascular region according to an embodiment of the present invention;

[0054] Figure 3 This is a comparative schematic diagram of normal blood vessels and overlapping blood vessels provided in one embodiment of the present invention;

[0055] Figure 4 This is a schematic diagram illustrating the representation of a distance factor according to an embodiment of the present invention;

[0056] Figure 5 This is a schematic diagram of the division of a region to be distinguished according to an embodiment of the present invention;

[0057] Figure 6 This is a partial schematic diagram of a blood vessel overlapping region provided in one embodiment of the present invention;

[0058] Figure 7 This invention provides a navigation path planning process according to one embodiment of the present invention. Detailed Implementation

[0059] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an auxiliary path planning method for inferior vena cava filter placement puncture target point proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0061] The following describes in detail, with reference to the accompanying drawings, a specific scheme of the auxiliary path planning method for inferior vena cava filter placement puncture target point provided by the present invention.

[0062] Please see Figure 1 The diagram illustrates a method flowchart for an auxiliary path planning method for inferior vena cava filter placement puncture target point according to an embodiment of the present invention. The method includes the following steps:

[0063] Step S1: Acquire multiple CT images of the patient's inferior vena cava and segment the vascular region in each CT image.

[0064] The placement of retrievable inferior vena cava (IVC) filters is commonly used to prevent and treat pulmonary embolism caused by deep vein thrombosis. A key aspect of the placement process is the precise identification of the puncture target, i.e., selecting a suitable puncture site and accurately placing the filter into the inferior vena cava. To improve the accuracy and safety of placement, pathway planning-assisted techniques have emerged. By combining modern imaging techniques (such as CT and ultrasound) with computer-aided navigation systems, doctors can be guided in real-time to accurately locate the puncture target during the procedure, assess the feasibility of the puncture path, and avoid damage to surrounding vital blood vessels and organs.

[0065] However, due to the complexity of vascular structures and the possibility of overlapping blood vessels, existing technologies that rely on the shortest distance between puncture target points to plan navigation paths cannot effectively avoid complex vascular structures. The presence of overlapping blood vessel areas can affect the accuracy of navigation paths. Therefore, in this embodiment of the invention, the main purpose is to analyze the impact of overlapping blood vessel areas on navigation path planning for puncture target points.

[0066] First, multiple CT images of the patient's inferior vena cava are needed. Specifically, a high-resolution CT scanner can be used to scan the area of ​​the inferior vena cava. During the CT scan, an iodine-containing contrast agent is injected to enhance the contrast between the blood vessel and the surrounding tissue. The CT scanner will acquire multiple CT images of the patient's inferior vena cava, each image showing different structures within the inferior vena cava, including blood vessels, organs, and bones.

[0067] It should be noted that during the scanning process, in this embodiment of the invention, a CT image is captured every second, and several CT images are obtained after the scan is completed.

[0068] Since the acquired CT images may contain various structures, and the analysis in this embodiment mainly focuses on the vascular region, it is necessary to segment the vascular region in each CT image.

[0069] Preferably, in one embodiment of the present invention, the method for obtaining the vascular region includes:

[0070] Each CT image is used as input to a pre-trained neural network, which outputs the vascular region within each CT image. When a blood vessel branches, it is recorded as two distinct vessels; that is, a vascular region may contain multiple blood vessels. Please refer to [link to relevant documentation]. Figure 2 This illustrates a schematic diagram of the division of the vascular region in one embodiment of the present invention.

[0071] It should be noted that the neural network in this embodiment of the present invention can be a CNN, and the training process of the neural network is a well-known technique, the specific process of which will not be described in detail here.

[0072] In the embodiments of the present invention, the acquisition and purchase of CT image data are authorized by the relevant users, and the process does not violate relevant laws and regulations, nor does it violate public order and good morals.

[0073] Step S2: In each CT image, based on the change in the distance between the two edge lines of each blood vessel in the vascular region, divide the region to be distinguished; analyze the shape characteristics of each region to be distinguished in each CT image and the area change between each region to be analyzed and the corresponding region in other CT images, and determine the vascular overlap characterization evaluation index of each region to be distinguished to distinguish the vascular overlap region and the non-vascular overlap region.

[0074] When planning the navigation path for the placement of a retrievable inferior vena cava filter, the inferior vena cava region often presents complex vascular structures and overlapping vessels. These overlapping vascular structures can result in multiple intertwined vascular images on CT scans. Since the accuracy of the navigation path directly affects the smooth placement of the inferior vena cava filter, the overlapping vessels not only increase the ambiguity of the navigation path but may also reduce the accuracy of puncture target confirmation. Therefore, it is necessary to distinguish the overlapping vascular areas in the CT images of the inferior vena cava during the path planning process for the placement of the retrievable inferior vena cava filter.

[0075] Blood vessel overlap can alter the morphology of vessels on CT scans. Normally, the diameter of a blood vessel gradually decreases from its tip to its tip; this change is usually smooth and uniform, ensuring blood flows at the expected pressure and velocity. In areas where there is no overlap, the vessel diameter remains normal. However, when vessels overlap, especially when two or more vessels overlap or cross at a certain point, the diameter changes locally. In the overlapping area, the vessel diameter widens significantly due to the overlap; the transition from the overlapping to the non-overlapping area becomes uneven, potentially exhibiting abrupt changes, resulting in a non-uniform increase or decrease in vessel diameter. (See also: [link to relevant documentation]). Figure 3 The illustration shows a comparative schematic diagram of normal blood vessels and overlapping blood vessels in one embodiment of the present invention. In CT images, the diameter change of blood vessels can be characterized by the change in the distance between the two edge lines of each blood vessel in the vascular region. Therefore, based on the aforementioned features, the vascular region is divided to obtain the region to be distinguished. At this time, the region to be distinguished may be the overlapping region of blood vessels or the non-overlapping region of blood vessels.

[0076] For overlapping vascular regions, in addition to significant changes in the shape of the vessels, the area of ​​the overlapping vessels will vary in different CT images. For non-overlapping vessels, only the area of ​​individual vessels differs, while the area of ​​a single vessel does not change significantly across different CT images. Therefore, based on the regions to be distinguished, the shape characteristics of the regions to be distinguished and the area changes between each region to be distinguished and its corresponding region in other CT images are considered to calculate a vascular overlap characterization evaluation index for each region to be distinguished. This index can more accurately reflect whether there is vascular overlap, and therefore, based on this index, it can be determined whether the region to be distinguished belongs to a vascular overlap region.

[0077] First, in each CT image, the region to be distinguished is divided based on the change in the distance between the two edge lines of each blood vessel in the vascular region. Preferably, in one embodiment of the present invention, the method for obtaining the region to be distinguished includes:

[0078] In each CT image, the edge line of each blood vessel in the vascular region can be obtained based on the Canny operator, and one of the two edge lines of each blood vessel can be selected as the target line.

[0079] On the target line, feature points are obtained based on the SURF feature point detection algorithm. The purpose of obtaining feature points is to simplify subsequent calculation steps, allowing for the segmentation of the region to be distinguished by analyzing only a few key locations. A straight line perpendicular to the target line is drawn through each feature point, and the intersection of this line and another edge line is used as the reference point for each feature point. The distance between each feature point and its corresponding reference point is then used as a distance factor. This distance factor characterizes the diameter of the blood vessel at each feature point, used to quantify the variation in blood vessel diameter. A larger distance factor indicates a larger blood vessel diameter, and vice versa. Please refer to [link to relevant documentation]. Figure 4 This illustrates a schematic diagram representing the distance factor in one embodiment of the present invention.

[0080] It should be noted that the Canny operator and the SURF feature point detection algorithm are both well-known technologies, and the specific process will not be elaborated here.

[0081] Then, the changes in the diameter of the blood vessels can be analyzed to divide the regions to be distinguished, so that the diameter of the blood vessels in each region to be distinguished has a relatively consistent change.

[0082] In this embodiment of the invention, the region to be distinguished is mainly divided by traversing feature points:

[0083] On the target line, a feature point that does not exist in the feature point set is randomly selected as the test point, and the test point corresponds to a feature point set.

[0084] Starting from the point to be measured, the feature points on both sides are traversed along the target line. For each feature point traversed, the difference in distance factor between the feature point and the feature points in the feature point set is analyzed to obtain the merging index. When the merging index meets the preset conditions, the traversed feature points are merged into the feature point set of the point to be measured, and the traversal continues; if the conditions are not met, the traversal stops.

[0085] The preset condition is set to the merge index being greater than or equal to the preset merge threshold.

[0086] On the target line, randomly select a feature point that does not exist in the feature point set as the test point and repeat the above traversal process until all feature points exist in the feature point set, and then stop to obtain the set of all feature points on the target line.

[0087] Methods for obtaining merged metrics include:

[0088] Each feature point encountered during the traversal is treated as a point to be analyzed.

[0089] The absolute value of the difference between the distance factor of the point to be analyzed and the distance factor of each feature point in the feature point set corresponding to the point to be analyzed is calculated as the difference factor. At this point, there is a difference factor between the point to be analyzed and each feature point in the feature point set. A larger difference factor indicates a greater difference between the diameter of the blood vessel at the point to be analyzed and the diameter of the blood vessel at the feature points in the feature point set. Therefore, if the mean of all difference factors is larger, the point to be analyzed should not be included in the feature point set. Conversely, a smaller difference factor indicates a higher consistency between the diameter of the blood vessel at the point to be analyzed and the diameter of the blood vessel at the feature points in the feature point set. Therefore, if the mean of all difference factors is smaller, the point to be analyzed should be included in the feature point set. Therefore, the mean of all difference factors corresponding to the point to be analyzed can be negatively correlated and normalized to correct the logical relationship and obtain a merging index. A larger merging index indicates that the point to be analyzed should be included in the feature point set. The negative correlation mapping and normalization here can be performed using the formula exp(-x), where exp() represents an exponential function with the natural constant e as the base, and x represents the independent variable.

[0090] Based on the aforementioned process, feature points on the target line of each blood vessel in the vascular region can be divided, so that continuous feature points with relatively consistent features exist in the same set of feature points. Finally, in each set of feature points, the two feature points that are farthest apart on the target line are taken as endpoints, and the area enclosed by the line segment between the endpoint and the corresponding reference point, the line segment of the reference point corresponding to the two endpoints on the edge line, and the line segment of the two endpoints on the target line is taken as a region to be distinguished.

[0091] It should be noted that the preset merging threshold is set to 0.65, and the specific value can be adjusted according to the implementation scenario, without limitation here.

[0092] Here's an example illustrating the process of obtaining the region to be distinguished: Initially, on the target line, all feature points are not in the feature point set. Therefore, feature point 10 is randomly selected as the test point, and feature point 10 corresponds to a feature point set a. At this point, feature point set a contains only feature point 10. Starting from the test point 10, the feature points on both sides are traversed along the target line. First, the traversal proceeds in one direction along the target line, and the feature point encountered is feature point 11. The merging index between feature point 11 and feature point 10 in feature point set a is calculated. If the merging index is greater than or equal to a preset merging threshold, feature point 11 is merged into feature point set a. At this point, feature point set a contains both feature point 10 and feature point 11. Then, feature point 12 is traversed. The merging index between feature point 12, feature point 10, and feature point 11 is calculated. If the merging index is still greater than or equal to the preset merging threshold, feature point 12 is also merged into feature point set a. At this point, feature point set a contains feature points 10 and 11. 12. Continue repeating the above process. If the preset condition is not met when traversing to feature point 15, stop. At this time, feature points 10, 11, 12, 13, and 14 exist in the feature point set. Then, starting from the test point 10, traverse along the target line in another direction. The specific process is the same as above. When the traversal in this direction is completed, the first feature point set a can be obtained. If feature points 7, 8, 9, 10, 11, 12, 13, and 14 exist in feature point set a, then you need to arbitrarily select a feature point from the other feature points besides these feature points as the test point and repeat the above process until all feature points on the target line exist in the feature point set. At this time, several feature point sets can be obtained. Taking feature point set a as an example, the region to be distinguished corresponding to feature point set a should consist of the line segments between feature points 7 and 14 on the target line, the line segment between feature point 7 and its corresponding reference point 7′, the line segment between feature point 14 and its corresponding reference point 14′, and the line segment between reference points 7′ and 14′ on the edge line. These four line segments enclose a region to be distinguished. Please refer to [link / reference]. Figure 5 This illustrates a schematic diagram of the division of regions to be distinguished in one embodiment of the present invention.

[0093] At this point, each blood vessel in the vascular region can be divided in each CT image, resulting in the region to be distinguished. The shape characteristics of the region to be distinguished and the area changes between the region to be distinguished and the corresponding region in other CT images can be further analyzed to determine the evaluation index of vascular overlap characterization for each region to be distinguished.

[0094] Preferably, in one embodiment of the present invention, the method for obtaining the evaluation indicators for vascular overlap includes:

[0095] Matching feature points in different CT images: Since the acquisition process of CT images follows certain parameters such as scanning interval and scanning direction, there are certain spatial and temporal relationships between different CT images. Therefore, the relative movement distance between feature points in each CT image and pixels in adjacent CT images can be calculated. This allows us to determine the corresponding position of feature points in each region to be distinguished in each CT image in other CT images, thereby obtaining the corresponding region. The corresponding region of each region to be distinguished in each CT image in other CT images is then used as the comparison region.

[0096] Since the area of ​​overlapping blood vessels may appear in different CT images, while the area of ​​the area where blood vessels do not overlap does not change significantly, the difference between the largest and smallest area of ​​each region to be distinguished and all corresponding comparison regions is used as the first overlap characterization evaluation factor for each region to be distinguished. The larger the first overlap characterization evaluation factor, the more significant the area change, and the larger the blood vessel overlap characterization evaluation index, the more likely it is to be a blood vessel overlap area.

[0097] Then, the shape characteristics of each region to be distinguished are analyzed. When blood vessels overlap, certain undulations will occur, which can be characterized by the curvature value. Therefore, in each region to be distinguished, the midpoint of the line segment formed by connecting each feature point to the corresponding reference point is obtained, and the line connecting all midpoints is taken as the center line. The least squares method can be used to fit the center line here. The more undulations appear on the center line, the higher the probability that the region to be distinguished belongs to the blood vessel overlapping region.

[0098] The mean curvature value at all midpoints along each centerline was calculated as the curvature degree value, while the variance of the curvature value at all midpoints along each centerline was calculated as the fluctuation degree value. The larger the curvature degree value and the larger the fluctuation degree value, the greater the shape fluctuation and change of the region to be distinguished. Therefore, the normalized product of the curvature degree value and the fluctuation degree value corresponding to each region to be distinguished was used as the second overlap characterization evaluation factor for each region to be distinguished. The second overlap characterization evaluation factor integrates the curvature degree and morphological stability characteristics of the region to be distinguished. The larger the second overlap characterization evaluation factor, the more likely the region to be distinguished is a vascular overlap region.

[0099] Finally, the sum of the first and second overlap characterization evaluation factors corresponding to each region to be distinguished is normalized and used as the vascular overlap characterization evaluation index for each region to be distinguished. A larger vascular overlap characterization evaluation index indicates that the area of ​​the region to be distinguished varies significantly across different CT images, and its morphology has undergone considerable fluctuations, thus making it more likely to be a vascular overlap region. Normalization is a technique well-known to those skilled in the art; the normalization function can be linear normalization or standard normalization, etc., and the specific normalization method is not limited here.

[0100] It should be noted that the use of the least squares method is a well-known technique, and the specific process will not be elaborated here.

[0101] Based on the aforementioned steps, an evaluation index for the vascular overlap characterization of each region to be distinguished in each CT image can be obtained. Therefore, it is possible to determine whether the region to be distinguished belongs to a vascular overlap region or a non-vascular overlap region based on this index.

[0102] Preferably, in one embodiment of the present invention, distinguishing between overlapping vascular regions and non-overlapping vascular regions includes:

[0103] Based on the above analysis, it can be seen that when the vascular overlap characterization evaluation index is larger, it indicates that the area to be distinguished is more likely to be a vascular overlap area. Therefore, an overlap threshold is preset. When the vascular overlap characterization evaluation index of a certain area to be distinguished is greater than the preset overlap threshold, the area to be distinguished is considered to be a vascular overlap area.

[0104] When the evaluation index of the vascular overlap characterization of a certain region to be distinguished is less than or equal to the preset overlap threshold, the region to be distinguished is considered to be a non-vascular overlap region.

[0105] It should be noted that in this embodiment of the present invention, the preset overlap threshold is 0.6, and the specific value can be adjusted according to the implementation scenario, and is not limited here.

[0106] At this point, the overlapping and non-overlapping vascular regions within the vascular area of ​​each CT image can be identified.

[0107] Step S3: In each CT image, each overlapping vascular region is combined with a non-overlapping vascular region to obtain multiple region combinations; within each region combination, the shape similarity between the overlapping vascular regions and the non-overlapping vascular regions, as well as the similarity between the vascular overlap characterization evaluation indicators of the non-overlapping vascular regions, are analyzed to determine the smooth continuation index of the region combination; the differences between the smooth continuation indices of all region combinations are analyzed, and combined with the proportion of overlapping vascular regions in the region combinations, matching combinations are selected from all region combinations.

[0108] When blood vessels overlap, the navigation path cannot be presented as clearly as expected. In other words, it is impossible to determine the matching between the overlapping part and the non-overlapping blood vessels on both sides. Therefore, it is necessary to distinguish which non-overlapping parts may belong to the same blood vessel as the overlapping part.

[0109] In this embodiment of the invention, firstly, in each CT image, each overlapping vascular region is combined with a non-overlapping vascular region to obtain multiple region combinations. Since the vascular edges between overlapping and non-overlapping portions on the same vessel usually exhibit a smoother connection, and the direction of vessel extension is consistent, specifically, the shape similarity between overlapping and non-overlapping vascular regions should be higher in each region combination. Simultaneously, the connection of the same vessel at the overlapping region may have high consistency. Therefore, in each region combination, the differences in vascular overlap characterization evaluation indicators between non-overlapping vascular regions are also considered, thereby calculating the smoothness continuation index for each region combination. This index characterizes the probability that the various parts of the region combination belong to the same vessel. Further, it is necessary to screen out matching combinations with a high degree of matching from all region combinations. During the screening process, since a larger proportion of overlapping vascular regions reduces the difficulty of matching and improves accuracy, the differences in the smoothness continuation index among all region combinations are analyzed, and the proportion of overlapping vascular regions in the region combinations is used as a weight to screen matching combinations from all region combinations.

[0110] First, in each CT image, a region combination is determined. Preferably, in one embodiment of the present invention, the method for obtaining the region combination includes:

[0111] In each CT image, for any overlapping vascular region, the non-overlapping vascular region connected to that overlapping vascular region is selected as a candidate region.

[0112] In the overlapping region of the blood vessels, select one candidate region in each of the two directions along the edge of the blood vessel to form a region combination with the overlapping region of the blood vessel, and obtain all non-repeating region combinations.

[0113] Here is an example illustrating the method for obtaining region combinations: If there are four non-vascular overlapping regions connected to a certain vascular overlapping region (1), two on each side, denoted as (2), (3), (4), and (5) respectively, please refer to [link to relevant documentation]. Figure 6The diagram shows a partial schematic of the overlapping region of blood vessels in one embodiment of the present invention. If (2), (3), (4), and (5) are taken as candidate regions of the overlapping region of blood vessels, then all non-repeating region combinations should be {(1), (2), (4)}, {(1), (2), (5)}, {(1), (3), (4)}, {(1), (3), (5)}.

[0114] At this point, we can obtain the region combination corresponding to each overlapping vascular region in each CT image. Then, in each region combination, we analyze the shape similarity between the overlapping vascular region and the non-overlapping vascular region, as well as the similarity between the vascular overlap characterization evaluation indicators of the non-overlapping vascular region, and determine the smooth continuation index of the region combination.

[0115] Preferably, in one embodiment of the present invention, the method for obtaining the smoothing continuation index includes:

[0116] Since the overlapping and non-overlapping parts of the same blood vessel usually exhibit a smoother connection at the vessel edges, and the extension direction of the blood vessel is consistent, in this embodiment of the invention, the consistency is mainly manifested in the smooth continuation of the center line of the two edge lines of the blood vessel. The center line can serve as the main axis of the blood vessel and can well reflect the extension and shape characteristics of the blood vessel.

[0117] Therefore, in each region combination, the intersection point between the centerline of each non-vascular overlapping region and the centerline of each vascular overlapping region is taken as the junction point. The junction point is the connection point of the main axes of different regions in the region combination, which helps to identify the branching and connection of blood vessels. The method for obtaining the centerline here is described in step S2. At this time, each non-vascular region corresponds to a junction point.

[0118] The centerline is fitted to all midpoints, allowing us to obtain the curvature value at each midpoint to reflect shape characteristics and aid in analyzing smoothness and continuity. The Euclidean distance between the midpoint of the centerline corresponding to the maximum curvature value and the corresponding intersection point of each non-vascular overlapping region is used as the positional deviation factor. The location of the maximum curvature value represents the position where the non-vascular overlapping region has the greatest curvature. Therefore, a smaller positional deviation factor indicates that the non-vascular overlapping region and the vascular overlapping region can maintain a good smooth connection at the position of greatest curvature; conversely, a larger positional deviation factor may indicate poor smoothness and continuity between the two regions at the intersection. The sum of the positional deviation factors corresponding to the two non-vascular overlapping regions in each region combination is negatively correlated to correct the logical relationship, thus obtaining a smoothness factor. A larger smoothness factor indicates better smoothness and continuity between the three regions in the region combination, and is more likely to be the same vessel. This negative correlation mapping can be achieved using the formula exp(-x), where exp() represents an exponential function with the natural constant e as the base, and x represents the independent variable.

[0119] If three regions in a region combination belong to the same blood vessel, then the similarity between the vascular overlap characterization evaluation indicators of the non-overlapping regions should be higher, reflecting a high degree of consistency in the connectivity between the three regions at the vascular overlap area. Therefore, in each region combination, the absolute value of the difference between the vascular overlap characterization evaluation indicators of two non-overlapping regions is calculated. The smaller the absolute value of this difference, the higher the similarity. This absolute value is then negatively correlated and normalized to correct the logical relationship, resulting in a similarity factor. The larger the similarity factor, the higher the consistency in the connectivity between the three regions, i.e., the higher the smoothness and continuity. This negative correlation mapping and normalization can be performed using the formula exp(-x), where exp() represents an exponential function with the natural constant e as the base, and x represents the independent variable.

[0120] Based on the foregoing analysis, the larger the smoothing factor and similarity factor of each region combination, the better the continuity among the three regions in the combination. Therefore, the sum of the smoothing factor and similarity factor corresponding to each region combination is normalized and used as the smoothing continuity index for each region combination. A larger smoothing continuity index for a region combination indicates that the three regions in that combination are more likely to belong to the same blood vessel, and is considered to have a higher degree of matching. Normalization is a technique well-known to those skilled in the art. The choice of normalization function can be linear normalization or standard normalization, etc., and the specific normalization method is not limited here.

[0121] Furthermore, based on the aforementioned process, it is known that each overlapping vascular region may correspond to multiple region combinations, but not all region combinations can correctly represent the matching situation between regions. Therefore, it is necessary to select the region combinations with a high degree of matching from all region combinations as the matching combinations.

[0122] If three regions in a region combination belong to the same blood vessel, then the smoothness continuation index of that region combination should be more prominent than that of other region combinations. At the same time, since the proportion of overlapping blood vessels is larger, the proportion of non-overlapping blood vessels is smaller, thus reducing the difficulty of matching and increasing the accuracy of matching. Therefore, the proportion of overlapping blood vessels in the region combination is used as a weight to screen matching combinations among all region combinations.

[0123] Preferably, in one embodiment of the present invention, the method for obtaining the matching combination includes:

[0124] In each CT image, for any overlapping blood vessel region, the proportion of the smoothness continuation index of each region combination corresponding to that overlapping region in the total smoothness continuation index of all region combinations corresponding to that overlapping region is calculated. In other words, the sum of the smoothness continuation indices of all region combinations corresponding to that overlapping region is used as the denominator, and the smoothness continuation index of each region combination is used as the numerator, thus obtaining the first matching contribution of each region combination. The larger the first matching contribution of a region combination, the more prominent its smoothness continuation is among all region combinations, and therefore the higher its matching degree.

[0125] Then, in each region combination corresponding to the overlapping blood vessel region, the average length of the two edge lines of the blood vessels in the overlapping region is calculated as the first length factor, and the average length of the two edge lines of the blood vessels in each non-overlapping region is calculated as the second length factor. The proportion of the first length factor in the sum of the first length factor and all second length factors is calculated. That is, the sum of the length factors corresponding to the three regions in the region combination (the sum of the first length factor and the two second length factors) is used as the denominator, and the first length factor is used as the numerator to obtain the second matching contribution of each region combination. The larger the second matching contribution of a region combination, the larger the proportion of the overlapping blood vessel region in that region combination, and the smaller the proportion of the non-overlapping blood vessel region. Therefore, the matching difficulty is lower and the matching accuracy is higher.

[0126] Finally, the sum of the first and second matching contributions for each region combination corresponding to the overlapping blood vessel region is normalized and used as the matching coefficient for each region combination. Based on the aforementioned analysis, the larger the matching coefficient, the higher the degree of matching between the three regions in the region combination, and the more likely they belong to the same blood vessel. Therefore, among all region combinations in the overlapping blood vessel region, the region combinations with matching coefficients greater than the preset matching threshold are considered as matching combinations. Each region in the matching combination is more likely to represent different parts of the same blood vessel. Normalization is a technique well-known to those skilled in the art, and the choice of normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0127] It should be noted that in this embodiment of the present invention, the preset matching threshold is 0.7, and the specific value can be adjusted according to the actual scenario, and is not limited here.

[0128] Step S4: In each CT image, analyze the difference in gray values ​​between non-vascular overlapping regions in the matching combination corresponding to each overlapping region of blood vessels, perform hierarchical evaluation on the non-vascular overlapping regions, and obtain the hierarchical evaluation results; plan the navigation path of the puncture target point according to the hierarchical evaluation results.

[0129] Based on the aforementioned steps, matching combinations corresponding to each overlapping vascular region can be obtained, and each part in each matching combination is more likely to belong to the same vascular vessel. In this embodiment of the invention, in order to further determine which vascular vessels are located in the upper layer and which are located in the lower layer, considering that the grayscale representation of vascular vessels located in different layers in CT images will be different, this feature is used to perform a layer assessment on the non-overlapping vascular regions in the matching combinations to obtain a layer assessment result. The layer assessment result can more accurately reflect which vascular vessels on both sides of the overlapping vascular region belong to the same layer. Therefore, when planning the navigation path of the puncture target point, for areas with overlapping vascular vessels, it can effectively avoid penetrating other vascular vessels and obtain a more accurate navigation path.

[0130] First, in each CT image, the difference in gray values ​​between non-vascular overlapping regions is analyzed in the matching combination corresponding to each overlapping region of blood vessels. The non-vascular overlapping regions are then subjected to hierarchical evaluation to obtain the hierarchical evaluation results.

[0131] Preferably, in one embodiment of the present invention, the method for obtaining the hierarchical evaluation result includes:

[0132] The purpose of this process is to analyze the depth differences represented by the non-vascular overlapping areas on both sides of the vascular overlapping area, thereby performing a hierarchical assessment, identifying non-vascular overlapping areas at the same level, and achieving a more accurate matching effect.

[0133] Therefore, in all matching combinations corresponding to each overlapping blood vessel region, in all non-overlapping blood vessel regions along each direction of the blood vessel, the average gray value of all pixels in each non-overlapping blood vessel region is calculated as the gray value feature of each non-overlapping blood vessel region.

[0134] Since blood vessels located in the upper layer are closer to the scanning device, their grayscale values ​​appear larger in CT images; conversely, blood vessels located in the lower layer are farther from the scanning device, their grayscale values ​​appear smaller in CT images. Therefore, all non-vascular overlapping areas in each direction are sorted in ascending order according to their grayscale feature values ​​to obtain a sorted sequence.

[0135] Finally, in the two sorting sequences corresponding to the overlapping vascular regions, the two non-overlapping vascular regions with the same index value belong to the same level of blood vessels.

[0136] After performing a hierarchical assessment of the non-vascular overlapping areas on both sides of the vascular overlapping region, a deeper understanding of the relative positions of the vessels can be obtained, effectively avoiding penetration of other vessels. Furthermore, before entering the vascular overlapping region, a planned path has already been established for the placement of the retrievable inferior vena cava filter in the non-vascular overlapping area. Then, based on the existing planned path, a puncture target point in a non-vascular overlapping region at the same level as the corresponding non-vascular overlapping area can be selected, thus completing the navigation path planning for the placement of the retrievable inferior vena cava filter in the presence of vascular overlap. Please refer to [link to relevant documentation]. Figure 7 This illustrates a navigation path planning process in one embodiment of the present invention.

[0137] In summary, when planning the navigation path for the inferior vena cava filter insertion target point, the possibility of overlapping blood vessels can affect the accuracy of the navigation path planning. Therefore, this embodiment of the invention mainly analyzes the impact of overlapping blood vessels. First, multiple CT images of the patient's inferior vena cava are acquired, and the vascular region is segmented. Since the morphology changes when blood vessels overlap, that is, the distance between the edge lines of the blood vessels changes significantly, the vascular region is initially divided based on this feature, resulting in regions to be distinguished. Furthermore, when blood vessels overlap, the area of ​​the overlap may show different changes in different CT images. Therefore, by combining the shape characteristics of each region to be analyzed and the area changes of the region to be analyzed with the corresponding regions in other CT images, a vascular overlap characterization evaluation index for the regions to be distinguished is determined, and the overlapping and non-overlapping vascular regions are distinguished based on this index. To determine which non-overlapping vascular regions on both sides of the overlapping vascular region may belong to the same blood vessel, this embodiment of the invention combines the overlapping and non-overlapping vascular regions, resulting in multiple region combinations. Given that regions belonging to the same blood vessel should exhibit relatively consistent and smooth spatial continuity and morphology, the shape similarity between overlapping and non-overlapping vascular regions, as well as the similarity between overlap characterization evaluation indicators of non-overlapping vascular regions, were analyzed within each region combination to determine the smoothness continuity index of the region combination. Since a larger proportion of overlapping vascular regions corresponds to a smaller proportion of non-overlapping vascular regions, thus reducing matching difficulty, their weight should be increased. Therefore, based on the differences in smoothness continuity indicators and the proportion of overlapping vascular regions, matching combinations were selected across all region combinations. Regions within each matching combination represent a higher probability of belonging to the same blood vessel. Furthermore, since blood vessels at different levels exhibit different grayscale values ​​in CT images, this invention analyzes the differences in grayscale values ​​between non-overlapping vascular regions for each matching combination corresponding to an overlapping vascular region. This allows for hierarchical evaluation of the non-overlapping vascular regions, yielding hierarchical evaluation results. These results can more accurately characterize which regions belong to the same blood vessel. Consequently, for regions with overlapping vascular regions, the navigation path for the puncture target can be planned more precisely based on the hierarchical evaluation results.

[0138] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0139] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for auxiliary path planning at the puncture target point for inferior vena cava filter placement, characterized in that, The method includes: Multiple CT images of the patient's inferior vena cava were acquired, and the vascular region was segmented in each CT image; In each CT image, the regions to be distinguished are divided based on the change in the distance between the two edge lines of each blood vessel in the vascular region; the shape features of each region to be distinguished in each CT image and the area change between each region to be analyzed and the corresponding region in other CT images are analyzed to determine the vascular overlap characterization evaluation index of each region to be distinguished, thereby distinguishing vascular overlap regions and non-vascular overlap regions. In each CT image, each overlapping vascular region is combined with a non-overlapping vascular region to obtain multiple region combinations. Within each region combination, the shape similarity between overlapping and non-overlapping vascular regions and the similarity between vascular overlap characterization evaluation indicators in non-overlapping vascular regions are analyzed to determine the smoothness continuation index of the region combination. The differences in the smoothness continuation index among all region combinations are analyzed, and combined with the proportion of overlapping vascular regions in the region combinations, matching combinations are selected from all region combinations. In each CT image, the difference in gray values ​​between non-vascular overlapping regions is analyzed in the matching combination corresponding to each overlapping region. The non-vascular overlapping regions are then subjected to hierarchical evaluation to obtain the hierarchical evaluation results. The navigation path for the puncture target is planned based on the hierarchical evaluation results.

2. The auxiliary path planning method for inferior vena cava filter placement puncture target point according to claim 1, characterized in that, The method for obtaining the region to be distinguished includes: In each CT image, select one of the two edge lines of each blood vessel as the target line. On the target line, feature points are obtained based on the SURF feature point detection algorithm. A straight line perpendicular to the target line is drawn through each feature point, and the intersection of the line with another edge line is used as the reference point corresponding to each feature point. The distance between each feature point and its corresponding reference point is used as the distance factor. On the target line, a feature point that does not exist in the feature point set is randomly selected as the test point, and the test point corresponds to a feature point set. Starting from the point to be measured, the feature points on both sides are traversed along the target line. For each feature point traversed, the difference in distance factor between the feature point and the feature points in the feature point set is analyzed to obtain the merging index. When the merging index meets the preset conditions, the traversed feature points are merged into the feature point set of the point to be measured, and the traversal continues; if the conditions are not met, the traversal stops. The preset condition is set to the merged index being greater than or equal to the preset merged threshold. On the target line, randomly select a feature point that does not exist in the feature point set as the test point and repeat the above traversal process until all feature points exist in the feature point set, and then stop to obtain the set of all feature points on the target line. In each set of feature points, the two feature points farthest apart on the target line are taken as endpoints. The area enclosed by the line segments between the endpoints and their corresponding reference points, the line segments of the reference points on the edge line, and the line segments of the two endpoints on the target line is taken as a region to be distinguished.

3. The auxiliary path planning method for inferior vena cava filter placement puncture target point according to claim 2, characterized in that, The methods for obtaining the merged indicators include: Each feature point encountered during the traversal is treated as a point to be analyzed. Calculate the absolute value of the difference between the distance factor of the point to be analyzed and the distance factor of each feature point in the set of feature points corresponding to the point to be measured, and use it as the difference factor; The mean values ​​of all differential factors corresponding to the points to be analyzed are negatively correlated and normalized, and then used as the merged index.

4. The auxiliary path planning method for inferior vena cava filter placement puncture target point according to claim 2, characterized in that, The methods for obtaining the evaluation indicators for vascular overlap include: Feature points in different CT images are matched to determine the corresponding region of each region to be distinguished in each CT image in other CT images, and the corresponding region of each region to be distinguished in each CT image in other CT images is used as the comparison region. In each region to be distinguished and in all corresponding comparison regions, the difference between the area of ​​the largest region and the area of ​​the smallest region is used as the first overlap characterization evaluation factor for each region to be distinguished. In each region to be distinguished, the line connecting the midpoint of the line segment between the feature point and the corresponding reference point is taken as the center line. The mean of the curvature values ​​at all midpoints on each center line is calculated as the curvature value, and the variance of the curvature values ​​at all midpoints on each center line is taken as the fluctuation value. The normalized value of the product of the curvature value and the fluctuation value corresponding to each region to be distinguished is used as the second overlapping characterization evaluation factor for each region to be distinguished. The normalized sum of the first and second overlapping characterization evaluation factors for each region to be distinguished is used as the vascular overlapping characterization evaluation index for each region to be distinguished.

5. The auxiliary path planning method for inferior vena cava filter placement puncture target point according to claim 1, characterized in that, The distinction between overlapping and non-overlapping vascular regions includes: When the evaluation index of the vascular overlap characterization of a certain region to be distinguished is greater than the preset overlap threshold, the region to be distinguished is considered to be a vascular overlap region. When the evaluation index of the vascular overlap characterization of a certain region to be distinguished is less than or equal to the preset overlap threshold, the region to be distinguished is considered to be a non-vascular overlap region.

6. The auxiliary path planning method for inferior vena cava filter placement puncture target point according to claim 4, characterized in that, The method for obtaining the region combination includes: In each CT image, for any overlapping blood vessel region, the non-overlapping blood vessel region connected to the overlapping blood vessel region is selected as a candidate region. In the overlapping region of the blood vessels, select one candidate region in each of the two directions along the edge of the blood vessel to form a region combination with the overlapping region of the blood vessel, and obtain all non-repeating region combinations.

7. The auxiliary path planning method for inferior vena cava filter placement puncture target point according to claim 6, characterized in that, The method for obtaining the smooth continuation index includes: In each region combination, the intersection point between the straight line containing the center line of each non-vascular overlapping region and the straight line containing the center line of the vascular overlapping region is taken as the intersection point; The Euclidean distance between the midpoint of the maximum curvature value on the centerline of each non-vascular overlapping region and the corresponding intersection point is used as the position deviation factor. The sum of the position deviation factors corresponding to two non-vascular overlapping regions is negatively correlated and mapped to the value of the smoothing factor. The absolute value of the difference between the vascular overlap characterization evaluation indicators of two non-vascular overlap regions is negatively correlated and normalized, and then used as the similarity factor. The normalized sum of the smoothing factor and similarity factor for each region combination is used as the smoothing continuation index for each region combination.

8. The auxiliary path planning method for inferior vena cava filter placement puncture target point according to claim 1, characterized in that, The method for obtaining the matching combination includes: In each CT image, for any overlapping blood vessel region, the proportion of the smooth continuation index of each region combination corresponding to the overlapping blood vessel region in the smooth continuation index of all region combinations corresponding to the overlapping blood vessel region is used as the first matching contribution. In each region combination corresponding to the overlapping blood vessel region, the average length of the two edge lines of the blood vessels in the overlapping blood vessel region is calculated as the first length factor. The average length of the two edge lines of the blood vessels in each non-overlapping blood vessel region is calculated as the second length factor. The proportion of the first length factor in the sum of the first length factor and all second length factors is used as the second matching contribution. The normalized sum of the first and second matching contributions for each region combination corresponding to the overlapping blood vessel region is used as the matching coefficient. Among all the regions in the overlapping blood vessel region, the regions with a matching coefficient greater than the preset matching threshold are selected as the matching combinations.

9. The auxiliary path planning method for inferior vena cava filter placement puncture target point according to claim 6, characterized in that, The methods for obtaining the hierarchical evaluation results include: In all matching combinations, in all non-vascular overlapping regions along each direction of the blood vessel in the overlapping region of blood vessels, the mean gray value of all pixels in each non-vascular overlapping region is calculated as the gray value feature value of each non-vascular overlapping region. All non-vascular overlapping regions are sorted in ascending order according to their gray value feature values ​​to obtain a sorted sequence. In the two sorting sequences corresponding to the overlapping vascular regions, the two non-overlapping vascular regions with the same index value belong to the same level of blood vessels.

10. The auxiliary path planning method for inferior vena cava filter placement puncture target point according to claim 1, characterized in that, The method for obtaining the vascular region includes: Each CT image is used as input to a pre-trained neural network, which then outputs the vascular region in each CT image.

Images