Liver resection path planning method and system based on three-dimensional simulation

Abstract

The invention relates to the field of medical images, and discloses a liver resection path planning method and system based on three-dimensional simulation, and the method comprises the steps: carrying out the processing of image Dicom data, and generating a liver three-dimensional model; reading three-dimensional coordinate values of point cloud data of a calculus module in the liver three-dimensional model, and converting the three-dimensional coordinate values into a plurality of sub-convex hull models; summarizing and deforming the irregular three-dimensional point cloud set and the sub-convex hull models to obtain a total convex hull model; drawing a total tangent line between the first viewpoint coordinate and the total convex hull model to obtain a liver resection pre-tangent line, and filtering to obtain a combined resection curved surface; and calculating dissection evaluation indexes in different resection directions, and comparing to obtain a dissection scheme of the optimal resection direction. According to the method, in consideration of coincidence of the combined resection curved surface and the blood vessel module, the dissection scheme in the optimal resection direction is screened out, the screened dissection scheme makes contact with the blood vessel as little as possible, and it is ensured that the bleeding state is effectively controlled in the liver resection operation.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging, and in particular to a liver resection path planning method and system based on three-dimensional simulation. Background Art

[0002] The treatment of hepatolithiasis is mainly surgical, and liver resection is the most effective treatment. Surgical methods include traditional open liver resection and laparoscopic liver resection. Due to the small incision and trauma area, laparoscopic liver resection is widely used in clinical practice. In laparoscopic liver resection, the laparoscope is first inserted into the body, the liver is partially removed, and then the extrahepatic or contralateral bile duct stones are treated. During the hemihepatectomy, attention should be paid to the protection of the middle hepatic vein, and bleeding should be controlled to avoid surgical failure due to excessive bleeding.

[0003] In order to improve the success rate of liver resection surgery, the existing technology first uses computer technology to perform individualized three-dimensional reconstruction of the patient based on the original CT and MRI data, accurately displaying the location, size, number and venous distribution of the stones, and performs diagnostic analysis and simulated surgery based on the individualized three-dimensional reconstruction model, thereby improving the success rate of the operation.

[0004] In the prior art, in liver resection surgery, the surgeon generally controls the direction and position of liver resection to ensure that as many stones as possible are removed. However, the distribution of blood vessels in the liver is uncertain. In actual operation, it is impossible to ensure that the currently selected liver resection direction can contact the blood vessels as little as possible, thereby ensuring effective control of the bleeding state. Summary of the invention

[0005] To this end, the present invention provides a liver resection path planning method and system based on three-dimensional simulation, which effectively solves the technical problem in the prior art that it is impossible to ensure that the currently selected liver resection direction can contact the blood vessels as little as possible, thereby failing to ensure effective control of the bleeding state.

[0006] In order to solve the above technical problems, the present invention specifically provides the following technical solutions: a liver resection path planning method based on three-dimensional simulation, comprising the following steps:

[0007] The Dicom image data obtained from the patient examination is imported into the Mimics reconstruction system, and the Dicom image data is processed by the Mimics reconstruction system to generate a three-dimensional liver model;

[0008] Reading the three-dimensional coordinate values ​​of the point cloud data of the stone module in the three-dimensional liver model, and converting the point cloud data into a plurality of sub-convex hull models, and removing at least some of the sub-convex hull models;

[0009] Based on the point cloud addition algorithm, point clouds are added between adjacent sub-convex hull models to obtain a number of irregular three-dimensional point cloud sets, and the total convex hull model is obtained by summarizing and deforming the irregular three-dimensional point cloud sets and sub-convex hull models;

[0010] The first viewpoint coordinates under different resection directions are customized, and the total tangent line is drawn between the first viewpoint coordinates and the total convex hull model to obtain the liver resection pre-cut line formed on the outer edge of the liver three-dimensional model, and the combined resection surface is obtained after filtering;

[0011] Based on the number of overlapping points between the combined resection surface and the vascular module in the three-dimensional liver model, the anatomical evaluation indexes under different resection directions were calculated, and the anatomical scheme with the optimal resection direction was obtained by comparison.

[0012] Further, based on the point cloud addition algorithm, point clouds are added between adjacent sub-convex hull models to obtain a number of irregular three-dimensional point cloud sets, including the following steps:

[0013] Arrange the sub-convex hull models according to the model positions, and calculate the centroids of every two adjacent sub-convex hull models according to the arrangement order;

[0014] Draw a connection line based on the centroid of the two convex hull models, and customize the coordinates of the second viewpoint in the direction of the extension line of the connection line;

[0015] Classifying and screening the points on the first sub-convex hull model to obtain a first tangent point set consisting of first tangent points, and classifying and screening the points on the second sub-convex hull model to obtain a second tangent point set consisting of second tangent points;

[0016] Connecting points between the first tangent point set and the second tangent point set in a one-to-one correspondence manner, and establishing an irregular three-dimensional point cloud;

[0017] Repeat the above steps to obtain an irregular three-dimensional point cloud between every two adjacent sub-convex hull models.

[0018] Further, taking a first tangent point a1 in the first tangent point set as a starting point, connecting all second tangent points in the second tangent point set with the first tangent point a1 to obtain a plurality of tangent point line segments, and selecting and retaining the tangent point line segment with the shortest length;

[0019] Taking the next first tangent point a2 in the first tangent point set as the starting point, repeat the above steps until all the first tangent points are constructed to obtain tangent line segments, and construct an irregular three-dimensional point cloud based on the tangent line segments.

[0020] Furthermore, the normal vector of each point on the convex hull model on the corresponding surface is calculated The first vector is composed of each point on the sub-convex hull model and the second viewpoint coordinates The normal vector and the first vector Dot product to get the dot product result;

[0021] Among them, if the dot product result corresponding to the point is greater than 0, the point is marked as the first facing viewpoint; if the dot product result corresponding to the point is equal to 0, the point is marked as the first tangent point or the second tangent point; if the dot product result corresponding to the point is less than 0, the point is marked as the first back viewpoint.

[0022] Further, a total tangent line is drawn between the first viewpoint coordinates and the total convex hull model to obtain a liver resection pre-cut line formed at the outer edge of the three-dimensional liver model, comprising the following steps:

[0023] Calculate the normal vector of each point on the total convex hull model on the corresponding surface The second vector is formed by each point on the total convex hull model and the first viewpoint coordinates The normal vector and the first vector Dot product to get the dot product result;

[0024] If the dot product result of the point is greater than 0, the point is marked as the second facing viewpoint; if the dot product result of the point is equal to 0, the point is marked as the third tangent point; if the dot product result of the point is less than 0, the point is marked as the second back viewpoint;

[0025] Draw a number of total tangent lines from the first viewpoint coordinates to the third tangent point, obtain the coincidence points of the total tangent lines and the surface point cloud of the liver three-dimensional model, obtain the liver anatomical point cloud set, perform curve fitting on the liver anatomical point cloud set, and obtain the liver resection pre-cut line.

[0026] Further, based on the total tangent between the first viewpoint coordinates and the third tangent point, the tangent convex hull surface of the point cloud corresponding to the total tangent is calculated, the tangent convex hull surface is used as the first interface, the point cloud on the surface of the three-dimensional liver model that is outside the first interface is filtered out, and the surface point cloud of the liver to be removed that is within the first interface is retained;

[0027] Filtering the tangent convex hull surface of the point cloud corresponding to the total tangent, taking the surface of the three-dimensional liver model as the second interface, filtering out the point cloud on the tangent convex hull surface outside the second interface, and retaining the incision edge point cloud within the second interface;

[0028] The second back-facing viewpoints are used to form a second back-facing viewpoint set, point clouds outside the second interface in the second back-facing viewpoint set are filtered out, and back point clouds within the second interface are retained;

[0029] The combined resection surface is obtained by combining the surface point cloud, the cutting edge point cloud and the back point cloud of the liver to be resected.

[0030] Furthermore, based on the number of overlapping points between the combined resection surface and the vascular module in the three-dimensional liver model, the anatomical evaluation indexes under different resection directions are calculated, and the anatomical scheme of the optimal resection direction is obtained by comparison, including the following steps:

[0031] Get the point cloud quantity y1 of the blood vessel module;

[0032] The number of overlapping points y2 between the combined resection surface and the vascular module in the three-dimensional liver model is calculated based on point cloud matching;

[0033] The anatomical evaluation index M is calculated based on the number of point clouds of the vascular module y1 and the number of overlapping points y2 of the combined resection surface and the vascular module in the liver three-dimensional model using the following formula:

[0034]

[0035] Based on the above steps, the anatomical evaluation index M of the anatomical schemes under different resection directions is calculated, and the anatomical scheme with the highest anatomical evaluation index M is selected by comparison.

[0036] Further, the three-dimensional coordinate values ​​of the point cloud data of the stone module in the three-dimensional liver model are converted into polygonal data, and a three-dimensional n-sided convex hull surface is generated based on the three-dimensional model convex hull algorithm to obtain a sub-convex hull model;

[0037] Wherein, after all the sub-convex hull models are obtained, at least part of the sub-convex hull models are eliminated, including the following steps:

[0038] The liver segmentation area is constructed based on the maximum cutting volume, the stone discreteness in the liver segmentation area is calculated, the aggregation parameter is calculated by the following formula based on the stone discreteness and the number of stones, and the liver segmentation area with the highest aggregation parameter is screened out, and other convex hull models outside the liver segmentation area are eliminated;

[0039] T=w 1 *N+w 2 *s;

[0040] Where T is the aggregation parameter, w 1 、w 2 are weights, N is the number of stones, and s is the stone dispersion.

[0041] Furthermore, after obtaining the total convex hull model, the normal of the convex hull surface in the total convex hull model is calculated, and the total convex hull model is expanded by x millimeters along the normal direction to obtain an optimized total convex hull model.

[0042] In order to solve the above technical problems, the present invention further provides the following technical solutions: a liver resection path planning system based on three-dimensional simulation, comprising:

[0043] A three-dimensional modeling module is used to receive the input image Dicom data and generate a three-dimensional model of the liver;

[0044] A point cloud processing module is connected to the three-dimensional modeling module for communication. The point cloud processing module obtains the three-dimensional coordinate values ​​of the point cloud data of the stone module in the three-dimensional liver model, and converts the point cloud data into a plurality of sub-convex hull models. Based on a point cloud addition algorithm, point clouds are added between adjacent sub-convex hull models to obtain a plurality of irregular three-dimensional point cloud sets, and the total convex hull model is obtained by processing the points.

[0045] The scheme generation module automatically generates the first viewpoint coordinates under different resection directions, performs data processing on the first viewpoint coordinates and the total convex hull model to obtain the liver resection pre-cut line formed on the outer edge of the liver three-dimensional model, and obtains the combined resection surface after filtering;

[0046] A calculation module is communicated with the scheme generation module, and the calculation module calculates the anatomical evaluation index of the anatomical scheme under different resection directions according to the number of overlapping points between the combined resection surface and the vascular module in the three-dimensional liver model, and compares and selects the anatomical scheme with the optimal resection direction.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] In the present invention, after data processing of the point cloud data of the stone module in the three-dimensional liver model, the total convex hull model of the convergent stone module is obtained, different resection directions are selected on the basis of the total convex hull model, and the liver resection pre-cutting line and the combined resection surface are obtained according to the analysis of different resection directions. Considering the overlap of the combined resection surface and the vascular module, the anatomical scheme with the optimal resection direction is screened out, so that the screened anatomical scheme contacts the blood vessels as little as possible, thereby ensuring effective control of the bleeding state during liver resection surgery. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0050] Figure 1 A flowchart of a liver resection path planning method based on three-dimensional simulation provided by an embodiment of the present invention;

[0051] Figure 2 A schematic diagram of a first tangent point and a second tangent point of adjacent sub-convex hull models in an embodiment of the present invention;

[0052] Figure 3Schematic diagram of an irregular three-dimensional point cloud of adjacent sub-convex hull models in an embodiment of the present invention.

[0053] The numbers in the figure represent the following:

[0054] 1- Convex hull model; 2. Centroid; 3. Extension line of connecting line; 4. Second viewpoint coordinates; 5. First tangent point; 6. Second tangent point; 7. Irregular three-dimensional point cloud. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] like Figure 1 As shown, the present invention provides a liver resection path planning method based on three-dimensional simulation, comprising the following steps:

[0057] The Dicom image data obtained from the patient examination is imported into the Mimics reconstruction system, and the Dicom image data is processed by the Mimics reconstruction system to generate a three-dimensional liver model;

[0058] Reading the three-dimensional coordinate values ​​of the point cloud data of the stone module in the three-dimensional liver model, and converting the point cloud data into a plurality of sub-convex hull models, and removing at least some of the sub-convex hull models;

[0059] Based on the point cloud addition algorithm, point clouds are added between adjacent sub-convex hull models to obtain a number of irregular three-dimensional point cloud sets, and the total convex hull model is obtained by summarizing and deforming the irregular three-dimensional point cloud sets and sub-convex hull models;

[0060] The first viewpoint coordinates under different resection directions are customized, and the total tangent line is drawn between the first viewpoint coordinates and the total convex hull model to obtain the liver resection pre-cut line formed on the outer edge of the liver three-dimensional model, and the combined resection surface is obtained after filtering;

[0061] Based on the number of overlapping points between the combined resection surface and the vascular module in the three-dimensional liver model, the anatomical evaluation indexes under different resection directions were calculated, and the anatomical scheme with the optimal resection direction was obtained by comparison.

[0062] In the present invention, after data processing of the point cloud data of the stone module in the three-dimensional liver model, the total convex hull model of the convergent stone module is obtained, different resection directions are selected on the basis of the total convex hull model, and the liver resection pre-cutting line and the combined resection surface are obtained according to the analysis of different resection directions. Considering the overlap of the combined resection surface and the vascular module, the anatomical scheme with the optimal resection direction is screened out, so that the screened anatomical scheme contacts the blood vessels as little as possible, thereby ensuring effective control of the bleeding state during liver resection surgery.

[0063] In the present invention, the Mimics reconstruction system is used to process the image Dicom data to generate a liver three-dimensional model. Specifically, the patient's preoperative CT and ultrasound examination image Dicom data are first imported into the Mimics reconstruction system. The Mimics reconstruction system processes the image Dicom data, such as adjusting the image grayscale value, rendering, threshold segmentation, etc., to generate a liver three-dimensional model. In addition to the liver body module, the liver three-dimensional model also includes a stone module, a blood vessel module, etc.

[0064] After obtaining the three-dimensional model of the liver, the three-dimensional coordinate values ​​of the point cloud data of the stone module in the three-dimensional liver model are converted into polygonal data, and a three-dimensional n-sided convex hull surface is generated based on the three-dimensional model convex hull algorithm to obtain a sub-convex hull model, which corresponds one-to-one to the stone module.

[0065] Since the distribution of the sub-convex hull model in the three-dimensional liver model is consistent with the stone module, the discreteness of its distribution is uneven and may be distributed in different areas of the liver, and it is impossible to remove every stone. Only stones in a certain area can be removed, and stones in other areas can be cleaned. Therefore, after all the sub-convex hull models are obtained, at least part of the sub-convex hull models are eliminated, which specifically includes the following steps:

[0066] The liver segmentation area is constructed based on the maximum cutting volume, and the stone discreteness in the liver segmentation area is calculated. The aggregation parameter is calculated by the following formula based on the stone discreteness and the number of stones, and the liver segmentation area with the highest aggregation parameter is selected, and other convex hull models outside the liver segmentation area are eliminated;

[0067] T=w 1 *N+w 2 *s;

[0068] Where T is the aggregation parameter, w 1 、w 2 are weights, N is the number of stones, and s is the stone dispersion.

[0069] In the above embodiments, the maximum cutting volume is first preset, and different liver segmentation areas are constructed at different liver positions based on the maximum cutting volume. The volume of each liver segmentation area is less than or equal to the maximum cutting volume. The stone discreteness and the number of stones in each liver segmentation area are calculated. The aggregation parameter of the liver segmentation area is calculated based on the stone discreteness and the number of stones in a liver segmentation area. The weight is adjusted according to the actual situation. The higher the convergence parameter, the more and denser the number of stones in the liver segmentation area. The liver segmentation area with the highest convergence parameter is screened out, and other sub-convex hull models outside the liver segmentation area are eliminated. The elimination step is mainly to facilitate the aggregation of the sub-convex hull models corresponding to the stone module in the liver segmentation area with the highest convergence parameter into a total convex hull model, and the anatomical scheme is analyzed for the total convex hull model. The stones corresponding to the stone module outside the liver segmentation area are cleaned by other non-resection methods.

[0070] After removing some of the sub-convex hull models, the retained sub-convex hull models are processed, and point clouds are added between adjacent sub-convex hull models based on a point cloud addition algorithm to obtain several irregular three-dimensional point cloud sets, including the following steps:

[0071] Arrange the sub-convex hull models according to the model positions, and calculate the centroids of every two adjacent sub-convex hull models according to the arrangement order;

[0072] Draw a connection line based on the centroid of the two convex hull models, and customize the coordinates of the second viewpoint in the direction of the extension line of the connection line;

[0073] Classify and filter the points on the first sub-convex hull model to obtain a first tangent point set consisting of first tangent points, and classify and filter the points on the second sub-convex hull model to obtain a second tangent point set consisting of second tangent points;

[0074] Connect points between the first tangent point set and the second tangent point set in a one-to-one correspondence manner, and establish an irregular three-dimensional point cloud;

[0075] Repeat the above steps to obtain an irregular three-dimensional point cloud between every two adjacent sub-convex hull models.

[0076] In the above steps, the points are connected in a one-to-one correspondence between the first tangent point set and the second tangent point set, and an irregular three-dimensional point cloud is established. Specifically, the following method is used:

[0077] A first tangent point a1 in the first tangent point set is used as the starting point, and all second tangent points in the second tangent point set are connected with the first tangent point a1 to obtain a number of tangent line segments, and the tangent line segment with the shortest length is selected and retained;

[0078] Taking the next first tangent point a2 in the first tangent point set as the starting point, repeat the above steps until all the first tangent points are constructed to obtain tangent line segments, and construct an irregular three-dimensional point cloud based on the tangent line segments.

[0079] The specific steps for classifying and screening the points on the convex hull model are as follows: Calculate the normal vector of each point on the convex hull model on the corresponding surface The first vector is composed of each point on the convex hull model and the second viewpoint coordinates The normal vector and the first vector Dot product to get the dot product result;

[0080] Among them, if the dot product result corresponding to the point is greater than 0, the point is marked as the first facing viewpoint; if the dot product result corresponding to the point is equal to 0, the point is marked as the first tangent point or the second tangent point; if the dot product result corresponding to the point is less than 0, the point is marked as the first back viewpoint.

[0081] like Figure 2 and Figure 3 As shown, taking two adjacent sub-convex hull models as an example, the centroid 2 of the two sub-convex hull models 1 is first calculated, a connecting line is drawn between the centroids, a second viewpoint coordinate 4 is marked on the extension line 3 of the connecting line, and the points on the first sub-convex hull model 1 are classified and screened according to the above steps based on the second viewpoint coordinate 4 to obtain a first tangent point set composed of a first tangent point 5, and the points on the second sub-convex hull model 1 are classified and screened according to the above steps based on the second viewpoint coordinate 4 to obtain a second tangent point set composed of a second tangent point 6, and then the first tangent point set and the second tangent point set are connected one by one before the first tangent point 5 and the second tangent point 6 to form a tangent line segment, and each tangent line segment satisfies: the condition that the length of the tangent line segment obtained by taking the first tangent point 5 as the starting point and the second tangent point 6 as the end point is the shortest, and an irregular three-dimensional point cloud 7 is gradually constructed between the first tangent point set and the second tangent point set in a point-to-line and line-to-surface manner.

[0082] Irregular three-dimensional point clouds can be obtained between each pair of adjacent sub-convex hull models in sequence according to the above steps. A preliminary total convex hull model is obtained by summarizing the irregular three-dimensional point cloud set and the sub-convex hull models. After obtaining the total convex hull model, the normal of the convex hull surface in the total convex hull model is calculated, and the total convex hull model is expanded by x millimeters along the normal direction to obtain the optimized total convex hull model. This process belongs to the deformation process of the total convex hull model, which can ensure that the edge of the total convex hull model is at least x millimeters away from the sub-convex hull models, and the combined resection surface is also at least x millimeters away from the stone, which effectively avoids direct resection of the stone during the cutting process.

[0083] After obtaining the total convex hull model, it is necessary to analyze the anatomical schemes under different resection directions. Under different resection directions, the first viewpoint coordinates are different. Therefore, different first viewpoint coordinates are set, and the corresponding liver resection pre-cutting lines and combined resection surfaces are obtained based on different viewpoint coordinates.

[0084] Specifically, drawing a total tangent line between the first viewpoint coordinates and the total convex hull model to obtain a liver resection pre-cut line formed on the outer edge of the three-dimensional liver model includes the following steps:

[0085] Calculate the normal vector of each point on the corresponding surface of the total convex hull model The second vector is composed of each point on the total convex hull model and the coordinates of the first viewpoint The normal vector and the first vector Dot product to get the dot product result;

[0086] If the dot product result of the point is greater than 0, the point is marked as the second facing viewpoint; if the dot product result of the point is equal to 0, the point is marked as the third tangent point; if the dot product result of the point is less than 0, the point is marked as the second back viewpoint;

[0087] Draw a number of total tangent lines from the first viewpoint coordinates to the third tangent point, obtain the coincidence points of the total tangent lines and the surface point cloud of the liver 3D model, obtain the liver anatomical point cloud set, perform curve fitting on the liver anatomical point cloud set, and obtain the liver resection pre-cut line.

[0088] Through the above steps, the third tangent point can be obtained based on the first viewpoint coordinates, and several total tangent lines can be drawn from the first viewpoint coordinates to the third tangent point. The set of coincident points between the total tangent line and the surface point cloud of the liver three-dimensional model is the liver anatomical point cloud set. Curve fitting is performed on it to obtain the liver resection pre-cutting line, which should be a circular closed curve.

[0089] After obtaining the pre-cut line for liver resection, continue to analyze the combined resection surface. The specific steps are as follows:

[0090] Based on the total tangent between the first viewpoint coordinates and the third tangent point, the tangent convex hull of the point cloud corresponding to the total tangent is calculated, the tangent convex hull is used as the first interface, the point cloud on the surface of the three-dimensional liver model that is outside the first interface is filtered out, and the point cloud on the surface of the liver to be removed that is within the first interface is retained;

[0091] The tangent convex hull surface of the point cloud corresponding to the total tangent is filtered, and the surface of the three-dimensional liver model is used as the second interface, and the point cloud on the tangent convex hull surface outside the second interface is filtered out, and the incision edge point cloud within the second interface is retained;

[0092] The second back-facing viewpoints are used to form a second back-facing viewpoint set, point clouds outside the second interface in the second back-facing viewpoint set are filtered out, and back point clouds within the second interface are retained;

[0093] The combined resection surface is obtained by combining the surface point cloud, the cutting edge point cloud and the back point cloud of the liver to be resected.

[0094] The first step is to use the total tangent to obtain the tangent convex hull surface, take the tangent convex hull surface as the first interface, filter out the point cloud on the surface of the liver three-dimensional model that is outside the first interface, and retain the point cloud on the surface of the liver to be removed that is within the first interface. This part is to obtain the partial point cloud on the liver surface of the combined resection surface through filtering;

[0095] The second part is to filter the tangent convex hull surface of the point cloud corresponding to the total tangent, take the surface of the three-dimensional liver model as the second interface, filter out the point cloud on the tangent convex hull surface outside the second interface, and retain the point cloud of the cutting edge within the second interface. This part is obtained by filtering the partial point cloud of the combined resection surface inside the liver, that is, the side surface point cloud of the resection edge during liver resection;

[0096] The third part is to filter out the point cloud outside the second interface in the second back viewpoint concentration, and retain the back point cloud within the second interface. This part obtains the point cloud of the back surface inside the liver in the total convex hull model through filtering, that is, the back point cloud of the liver area to be removed;

[0097] The entire combined resection surface is obtained by summarizing the above three parts.

[0098] After the above steps, the pre-cutting lines and combined resection surfaces of liver resection under different resection directions can be obtained, which are the anatomical plans under different resection directions. For different anatomical plans, it is necessary to evaluate and analyze them considering whether the bleeding effect is easy to control.

[0099] Specifically, based on the number of overlapping points between the combined resection surface and the vascular module in the three-dimensional liver model, the anatomical evaluation indexes under different resection directions are calculated, and the anatomical scheme of the optimal resection direction is obtained by comparison, including the following steps:

[0100] Get the point cloud quantity y1 of the blood vessel module;

[0101] The number of overlapping points y2 between the combined resection surface and the vascular module in the three-dimensional liver model is calculated based on point cloud matching;

[0102] The anatomical evaluation index M is calculated based on the number of point clouds of the vascular module y1 and the number of overlapping points y2 of the combined resection surface and the vascular module in the liver three-dimensional model using the following formula:

[0103]

[0104] Based on the above steps, the anatomical evaluation index M of the anatomical schemes under different resection directions is calculated, and the anatomical scheme with the highest anatomical evaluation index M is selected by comparison.

[0105] In the above evaluation and analysis steps, the overlap between different combined resection surfaces and vascular modules is taken into consideration. Based on the above evaluation steps, the anatomical scheme with the least overlap can be screened out, and then the combined resection surface, liver resection pre-cutting line and the corresponding first viewpoint coordinates in the anatomical scheme are output. The combined resection surface and liver resection pre-cutting line can provide guidance for the resection position and resection path in liver resection surgery. The first viewpoint coordinates can be converted into the resection direction, providing guidance for the insertion direction of the laparoscope before liver resection and the insertion path of the resection structure.

[0106] The present invention also provides a system for a liver resection path planning method based on three-dimensional simulation, which comprises:

[0107] A three-dimensional modeling module is used to receive the input image Dicom data and generate a three-dimensional model of the liver;

[0108] The point cloud processing module is connected to the three-dimensional modeling module in communication. The point cloud processing module obtains the three-dimensional coordinate value of the point cloud data of the stone module in the three-dimensional liver model, and converts it into a number of sub-convex hull models. Based on the point cloud addition algorithm, point clouds are added between adjacent sub-convex hull models to obtain a number of irregular three-dimensional point cloud sets, and the total convex hull model is obtained by processing.

[0109] The scheme generation module automatically generates the first viewpoint coordinates under different resection directions, performs data processing on the first viewpoint coordinates and the total convex hull model to obtain the liver resection pre-cut line formed on the outer edge of the liver three-dimensional model, and obtains the combined resection surface after filtering;

[0110] The calculation module is connected to the scheme generation module in communication. The calculation module calculates the anatomical evaluation index of the anatomical scheme under different resection directions according to the number of coincidence points between the combined resection surface and the vascular module in the three-dimensional liver model, and compares and selects the anatomical scheme with the optimal resection direction.

[0111] The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application. The protection scope of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and protection scope of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present application.

Claims

1. A liver resection path planning method based on three-dimensional simulation, characterized in that: The following steps are involved: The Dicom image data obtained from the patient examination is imported into the Mimics reconstruction system, and the Dicom image data is processed by the Mimics reconstruction system to generate a three-dimensional liver model; Reading the three-dimensional coordinate values ​​of the point cloud data of the stone module in the three-dimensional liver model, and converting the point cloud data into a plurality of sub-convex hull models, and removing at least some of the sub-convex hull models; Based on the point cloud addition algorithm, point clouds are added between adjacent sub-convex hull models to obtain a number of irregular three-dimensional point cloud sets, and the total convex hull model is obtained by summarizing and deforming the irregular three-dimensional point cloud sets and sub-convex hull models; The first viewpoint coordinates under different resection directions are customized, and the total tangent line is drawn between the first viewpoint coordinates and the total convex hull model to obtain the liver resection pre-cut line formed on the outer edge of the liver three-dimensional model, and the combined resection surface is obtained after filtering; Based on the number of overlapping points between the combined resection surface and the vascular module in the three-dimensional liver model, the anatomical evaluation indexes under different resection directions were calculated, and the anatomical scheme with the optimal resection direction was obtained by comparison.

2. The method for liver resection path planning based on three-dimensional simulation according to claim 1, characterized in that: Based on the point cloud addition algorithm, point clouds are added between adjacent sub-convex hull models to obtain several irregular three-dimensional point cloud sets, including the following steps: Arrange the sub-convex hull models according to the model positions, and calculate the centroids of every two adjacent sub-convex hull models according to the arrangement order; Draw a connection line based on the centroid of the two convex hull models, and customize the coordinates of the second viewpoint in the direction of the extension line of the connection line; Classifying and screening the points on the first sub-convex hull model to obtain a first tangent point set consisting of first tangent points, and classifying and screening the points on the second sub-convex hull model to obtain a second tangent point set consisting of second tangent points; Connecting points between the first tangent point set and the second tangent point set in a one-to-one correspondence manner, and establishing an irregular three-dimensional point cloud; Repeat the above steps to obtain an irregular three-dimensional point cloud between every two adjacent sub-convex hull models.

3. The method for liver resection path planning based on three-dimensional simulation according to claim 2, characterized in that: Taking a first tangent point a1 in the first tangent point set as the starting point, connecting all second tangent points in the second tangent point set with the first tangent point a1 to obtain a number of tangent point segments, and selecting the tangent point segment with the shortest length and retaining it; Taking the next first tangent point a2 in the first tangent point set as the starting point, repeat the above steps until all the first tangent points are constructed to obtain tangent line segments, and construct an irregular three-dimensional point cloud based on the tangent line segments.

4. The method for liver resection path planning based on three-dimensional simulation according to claim 3, characterized in that: Calculate the normal vector of each point on the convex hull model on the corresponding surface The first vector is composed of each point on the sub-convex hull model and the second viewpoint coordinates The normal vector and the first vector Dot product to get the dot product result; Among them, if the dot product result corresponding to the point is greater than 0, the point is marked as the first facing viewpoint; if the dot product result corresponding to the point is equal to 0, the point is marked as the first tangent point or the second tangent point; if the dot product result corresponding to the point is less than 0, the point is marked as the first back viewpoint.

5. The method for liver resection path planning based on three-dimensional simulation according to claim 1, characterized in that: Drawing a total tangent line between the first viewpoint coordinates and the total convex hull model to obtain a liver resection pre-cut line formed on the outer edge of the three-dimensional liver model includes the following steps: Calculate the normal vector of each point on the total convex hull model on the corresponding surface The second vector is formed by each point on the total convex hull model and the first viewpoint coordinates The normal vector and the first vector Dot product to get the dot product result; If the dot product result of the point is greater than 0, the point is marked as the second facing viewpoint; if the dot product result of the point is equal to 0, the point is marked as the third tangent point; if the dot product result of the point is less than 0, the point is marked as the second back viewpoint; Draw a number of total tangent lines from the first viewpoint coordinates to the third tangent point, obtain the coincidence points of the total tangent lines and the surface point cloud of the liver three-dimensional model, obtain the liver anatomical point cloud set, perform curve fitting on the liver anatomical point cloud set, and obtain the liver resection pre-cut line.

6. The method for liver resection path planning based on three-dimensional simulation according to claim 5, characterized in that: Based on the total tangent between the first viewpoint coordinates and the third tangent point, the tangent convex hull of the point cloud corresponding to the total tangent is calculated, the tangent convex hull is used as the first interface, the point cloud on the surface of the three-dimensional liver model that is outside the first interface is filtered out, and the point cloud on the surface of the liver to be removed that is within the first interface is retained; Filtering the tangent convex hull surface of the point cloud corresponding to the total tangent, taking the surface of the three-dimensional liver model as the second interface, filtering out the point cloud on the tangent convex hull surface outside the second interface, and retaining the incision edge point cloud within the second interface; The second back-facing viewpoints are used to form a second back-facing viewpoint set, point clouds outside the second interface in the second back-facing viewpoint set are filtered out, and back point clouds within the second interface are retained; The combined resection surface is obtained by combining the surface point cloud, the cutting edge point cloud and the back point cloud of the liver to be resected.

7. The method for liver resection path planning based on three-dimensional simulation according to claim 1, characterized in that: Based on the number of overlapping points between the combined resection surface and the vascular module in the liver three-dimensional model, the anatomical evaluation indexes under different resection directions are calculated, and the anatomical scheme of the optimal resection direction is obtained by comparison, including the following steps: Get the point cloud quantity y1 of the blood vessel module; The number of overlapping points y2 between the combined resection surface and the vascular module in the three-dimensional liver model is calculated based on point cloud matching; The anatomical evaluation index M is calculated based on the number of point clouds of the vascular module y1 and the number of overlapping points y2 of the combined resection surface and the vascular module in the liver three-dimensional model using the following formula: Based on the above steps, the anatomical evaluation index M of the anatomical schemes under different resection directions is calculated, and the anatomical scheme with the highest anatomical evaluation index M is selected by comparison.

8. The method for liver resection path planning based on three-dimensional simulation according to claim 1, characterized in that: The three-dimensional coordinate values ​​of the point cloud data of the stone module in the three-dimensional liver model are converted into polygonal data, and a three-dimensional n-sided convex hull surface is generated based on the three-dimensional model convex hull algorithm to obtain a convex hull model; Wherein, after all the sub-convex hull models are obtained, at least part of the sub-convex hull models are eliminated, including the following steps: The liver segmentation area is constructed based on the maximum cutting volume, the stone discreteness in the liver segmentation area is calculated, the aggregation parameter is calculated by the following formula based on the stone discreteness and the number of stones, and the liver segmentation area with the highest aggregation parameter is screened out, and other convex hull models outside the liver segmentation area are eliminated; T = w1*N+w2*s; Where T is the aggregation parameter, w1 and w2 are weights, N is the number of stones, and s is the stone dispersion.

9. The method for liver resection path planning based on three-dimensional simulation according to claim 1, characterized in that: After obtaining the total convex hull model, the normal of the convex hull surface in the total convex hull model is calculated, and the total convex hull model is expanded by x millimeters along the normal direction to obtain the optimized total convex hull model.

10. A system using the liver resection path planning method based on three-dimensional simulation according to any one of claims 1 to 9, characterized in that: have: A three-dimensional modeling module is used to receive the input image Dicom data and generate a three-dimensional model of the liver; A point cloud processing module is connected to the three-dimensional modeling module for communication. The point cloud processing module obtains the three-dimensional coordinate values ​​of the point cloud data of the stone module in the three-dimensional liver model, and converts the point cloud data into a plurality of sub-convex hull models. Based on a point cloud addition algorithm, point clouds are added between adjacent sub-convex hull models to obtain a plurality of irregular three-dimensional point cloud sets, and the total convex hull model is obtained by processing the points. The scheme generation module automatically generates the first viewpoint coordinates under different resection directions, performs data processing on the first viewpoint coordinates and the total convex hull model to obtain the liver resection pre-cut line formed on the outer edge of the liver three-dimensional model, and obtains the combined resection surface after filtering; A calculation module is communicated with the scheme generation module, and the calculation module calculates the anatomical evaluation index of the anatomical scheme under different resection directions according to the number of overlapping points between the combined resection surface and the vascular module in the three-dimensional liver model, and compares and selects the anatomical scheme with the optimal resection direction.

Images