Method and device for positioning a surgical incision based on spatial voxelization judgment of ct scan data

Abstract

The application provides a CT scanning data operation incision positioning method and device based on spatial voxelization judgment, three-dimensional modeling is performed by using known CT data, a lesion point is marked first, organ segmentation and body surface extraction are performed, and a voxelized three-dimensional model is constructed, which serves as a basis for spatial analysis of incision positioning. The method for incision positioning is that an inverted cone is generated according to the lesion point as a conical apex, candidate incision points are searched in the range interval of the cone, path simulation is performed by using surgical operation, spatial analysis and collision detection are performed on the path, 2 to 3 candidate positions of surgical incisions are generated, and visualization is realized in a three-dimensional view. The application overcomes the problems that the current lung minimally invasive operation hole opening strategy depends on the personal experience of doctors, the strategy is not standard, and the operation positioning risk is high, and provides a quantifiable and analyzable operation incision positioning method for minimally invasive operation, and the scientificity and analyzability of the minimally invasive operation are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the combination of remote sensing image processing and spatial analysis in the medical field, and in particular to a CT scan data surgical incision positioning method and device based on spatial voxelization judgment. BACKGROUND

[0002] Pulmonary minimally invasive surgery detects the size and specific location of the tumor through chest CT scanning, and achieves the purpose of removing the pulmonary lesion through a small surgical incision (commonly known as an opening) and special surgical instruments. Minimally invasive surgery can significantly reduce surgical trauma, reduce postoperative pain, reduce surgical complications, promote rapid recovery of patients, and improve patient satisfaction.

[0003] The surgical path of the current pulmonary minimally invasive surgery is planned manually by the doctor observing the CT image of the patient according to experience, and is severely dependent on the personal experience of the doctor. Due to the single mode of the surgical incision, the incision position of the minimally invasive surgery cannot be optimal, and if the preoperative strategy is not ideal, it will greatly affect the success rate of the surgery, the size of the incision, and the use and control of the machine, causing tissue injury to the patient, increasing the pain of the patient, and prolonging the recovery time after the operation. SUMMARY

[0004] The present application mainly solves the technical problem that the existing minimally invasive surgical incision positioning is severely dependent on the personal experience and skills of the doctor, lacks quantitative analysis technology, and results in poor incision positioning effect. Therefore, the present application proposes a CT scan data surgical incision positioning method based on spatial voxel judgment. First, a voxelized three-dimensional model is established using CT scan data, and the model is subjected to organ segmentation, skin surface extraction and lesion point marking. Then, search positioning and optimization analysis are performed in the operating cone with the lesion point as the vertex, and the positioning of the surgical incision is completed. The core of the present application is a three-dimensional model reconstructed from CT scan data for spatial calculation and analysis, and a refined surgical incision positioning method based on spatial voxel judgment.

[0005] The method of the present application is based on CT scan data, uses a moving cube algorithm for three-dimensional modeling, voxelizes the three-dimensional model using a ray-polygon intersection test method, and obtains the best incision position through feasibility analysis and weight addition. It can be used in various scenarios of pulmonary minimally invasive incision surgery. The specific technical scheme adopted is as follows:

[0006] The first aspect discloses a CT scan data surgical incision positioning method based on spatial voxelization judgment, comprising:

[0007] Obtaining CT data of a patient and preprocessing the data;

[0008] Three-dimensional modeling and lesion point marking are performed on the preprocessed CT data;

[0009] performing organ segmentation and body surface contour extraction based on the constructed three-dimensional model;

[0010] voxelizing the three-dimensional model after performing organ segmentation and body surface contour extraction using a ray polygon intersection test method;

[0011] Based on the obtained three-dimensional voxel model, a reverse cone is generated with the lesion point center as the vertex, the body surface points covered by the cone are traversed respectively, the line connecting the body surface point and the lesion point center is taken as the surgical path, the feasibility of the surgical path is judged according to the important organ avoidance and the entry angle, the feasible skin points are screened out, and then the candidate incision position is screened out from all the feasible skin points according to the preset condition;

[0012] According to the three-dimensional voxel model and the candidate incision position, the target incision position is obtained.

[0013] In one embodiment, the preprocessed CT data is three-dimensionally modeled and the lesion point is marked, including:

[0014] The preprocessed CT data is read into the program, and the CT data is converted into HU values, which are used to represent the density of tissues in CT scanning;

[0015] The size of each voxel read in is resampled, and the HU values of the resampled image are interpolated using a trilinear interpolation method;

[0016] The interpolated image is modeled using a moving cube algorithm, specifically: according to the target setting of the threshold required for each organ to be extracted, points greater than or equal to the threshold required for each organ to be extracted are stored, thereby retaining the extracted organ data and forming several triangular faces on the cube, and connecting the triangular faces of each voxel; then draw a polygon, set the polygon color, add the polygon to the three-dimensional coordinates, and complete the creation of the three-dimensional model;

[0017] In the three-dimensional model, the coordinate points of the lesion point are traversed, and the threshold corresponding to the coordinate points of the lesion point is increased by 100.

[0018] In one embodiment, the way of converting CT data into HU values is:

[0019] HU = val × slope + intercept

[0020] val represents the original pixel value in the CT image, slope is the slope parameter of the CT scanner, and intercept represents the intercept parameter of the CT scanner.

[0021] In one embodiment, organ segmentation and body surface contour extraction are performed based on the constructed three-dimensional model, including:

[0022] The CT data is subjected to HU value conversion and resampling processing; the HU value is mapped to the interval of 0-255, and then the threshold is determined by using the maximum inter-class variance method; the contour is found based on the determined threshold, specifically: when there are both value and 0 points around a point, the point is saved and saved in a newly created array; the newly created array is subjected to 3D modeling to realize contour extraction;

[0023] The threshold required for separating the lung is set to -600, the image is binarized, and each region is labeled; the number of pixels contained in each connected region is counted according to the label; after retaining two largest connected regions, the holes are filled by using morphological method to obtain a lung model; then, a suitable seed point is selected, and a region growing method is used to obtain three-dimensional models of the lung organ, cardiovascular system and heart to complete organ segmentation.

[0024] In an embodiment, the three-dimensional model after organ segmentation and contour extraction of the body surface is voxelized using a ray polygon intersection test method, including:

[0025] A three-dimensional space composed of a plurality of voxels is established;

[0026] For the surface of the three-dimensional model, a projection to two-dimensional method is used to find the maximum and minimum values of the triangular network model in x, y and z directions respectively, and then a cube with edges parallel to the coordinate axes is established with the maximum and minimum values as the opposite vertices, as the AABB bounding box of the triangular network model; the intersection operation is performed on the AABB bounding box and the triangular network model to obtain the intersected voxel units of the triangular surface, i.e. the surface of the three-dimensional model;

[0027] For the interior of the three-dimensional model, all empty voxels in the AABB bounding box are traversed, and two rays are emitted from the center position in the axis-aligned direction; the intersection positions of the two rays with the triangular surface are calculated, and the normal vectors of the intersection points are obtained; then, it is judged whether the voxel is located outside or inside the three-dimensional model according to the relationship between the normal vectors of the intersection points.

[0028] In an embodiment, the preset conditions include the distance between the surgical path extension line and the important organ, the entry angle of the surgical needle and the skin surface, and the length of the lung puncture path.

[0029] In an embodiment, candidate incision positions are selected from all feasible skin points according to the preset conditions, including:

[0030] The collision probability of the surgical motion simulation is calculated according to the distance between the center of the voxelized model of the surgical needle and the important organ, and a collision risk sub-objective function is set;

[0031] The entry angle risk sub-objective function is set according to the skin entry angle and the lung parenchyma entry angle;

[0032] Set the path risk sub-objective function under the conditions of total puncture path length and intrapulmonary length;

[0033] By setting the weights of each risk sub-objective function, the total risk function at the cut location is obtained;

[0034] The weights of each skin point are calculated based on the total risk function of the incision location, and the three locations with the lowest risk are selected as candidate incision locations.

[0035] Based on the same inventive concept, a second aspect of the present invention discloses a surgical incision positioning device based on CT scan data and spatial voxelization determination, comprising:

[0036] The data acquisition module is used to acquire and preprocess the patient's CT data;

[0037] The 3D modeling module is used to perform 3D modeling and lesion point marking on preprocessed CT data;

[0038] The segmentation and extraction module is used for organ segmentation and body surface contour extraction based on the constructed 3D model;

[0039] The voxelization module is used to voxelize the 3D model after organ segmentation and body surface contour extraction using the ray polygon intersection test method.

[0040] The candidate incision location screening module is used to generate an inverted cone with the center of the lesion point as the vertex based on the obtained three-dimensional voxel model, traverse the body surface points covered by the cone, take the line connecting the body surface points and the center of the lesion point as the surgical path, judge the feasibility of the surgical path based on the avoidance of important organs and the puncture angle, screen out feasible skin points, and then screen out candidate incision locations from all feasible skin points according to preset conditions.

[0041] The surgical incision localization module is used to obtain the target incision location based on the three-dimensional voxel model and candidate incision locations.

[0042] Based on the same inventive concept, a third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the surgical incision localization method based on spatial voxelization judgment of CT scan data as described in the first aspect.

[0043] Based on the same inventive concept, the fourth aspect of the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the surgical incision localization method based on spatial voxelization judgment of CT scan data as described in the first aspect.

[0044] Compared with the prior art, the application has the following advantages and beneficial technical effects:

[0045] The application provides a CT scanning data surgical incision positioning method based on spatial voxelization judgment, which comprises the following steps: after three-dimensional modeling and lesion point labeling are performed on preprocessed CT data, organ segmentation and body surface contour extraction are performed based on the constructed three-dimensional model; then, voxelization is performed on the three-dimensional model after organ segmentation and body surface contour extraction by using a ray polygon intersection test method; then, a reverse cone is generated with the center of the lesion point as the vertex based on the obtained three-dimensional voxel model, and the body surface points covered by the reverse cone are traversed respectively, the line connecting the body surface point and the center of the lesion point is taken as a surgical path, the feasibility of the surgical path is judged according to important organ avoidance and puncture angle, and then the feasible skin points are screened out, and then candidate incision positions are screened out from all the feasible skin points according to preset conditions; finally, the target incision position is obtained according to the three-dimensional voxel model and the candidate incision position. The organ segmentation and body surface contour extraction are performed based on the constructed three-dimensional model, so that the three-dimensional modeling can be ensured to be fast. Then, the model voxelization is performed on the model output from the software, the voxel is a combination of pixel, volume and element, and is equivalent to a pixel in a 3D space. The model voxelization can realize the conversion of the organ model from a complex mesh to a uniform voxel, and the point cloud data is more orderly stored in the memory. Therefore, compared with the ordinary model, the voxelization can better adapt to the irregularly shaped model, can make the spatial convolution more efficient, is beneficial to extracting local information features of multiple levels and multiple scales, such as lung lesion and other tissue positions, and thus provides convenience for incision positioning. Moreover, the voxelization can convert the object into a volume data representation, and is easy to perform operations such as lesion volume calculation and incision positioning. When the surgical incision positioning is performed, the opening direction and range are determined by taking the length of the puncture instrument, the obstacle avoidance of important tissues and organs (such as chest wall tissues and blood vessels), the skin puncture angle, the visibility constraint and other constraint conditions as hard constraint conditions, and then the most suitable surgical incision point is screened out according to the distance between the surgical path extension line and the important organs, the collision probability of the surgical motion simulation and the length of the puncture path in the lung. The technical problem that the existing minimally invasive surgical incision positioning seriously depends on the personal experience and skills of doctors, the strategy lacks quantification and standardization, and the positioning effect is poor is solved. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0047] Figure 1 The whole flow chart of the CT scanning data surgical incision positioning method based on spatial voxelization judgment in the embodiment of the application is shown in the figure;

[0048] Figure 2 The patient coordinate system diagram in the embodiment of the application is shown in the figure;

[0049] Figure 3 The generation diagram of a triangular face in the moving cube algorithm in the embodiment of the application is shown in the figure;

[0050] Figure 4 The result example diagram of organ contour extraction in the embodiment of the application is shown in the figure;

[0051] Figure 5 The diagram of morphological operation inflation and corrosion in the embodiment of the application is shown in the figure;

[0052] Figure 6 The diagram of the ray polygon intersection test method in the embodiment of the application is shown in the figure;

[0053] Figure 7 The result example diagram of voxelization in the embodiment of the application is shown in the figure;

[0054] Figure 8 The process diagram of finding the best incision position in the embodiment of the application is shown in the figure. DETAILED DESCRIPTION

[0055] The application discloses a CT scanning data surgical incision positioning method based on spatial voxels, which utilizes known CT data to perform three-dimensional modeling, firstly marks a lesion point, performs organ segmentation and body surface extraction, and then takes the organ segmentation result as the basis for spatial analysis of incision positioning. An inverted cone is generated with the lesion point as the cone vertex, and the incision point is searched within the cone range. A surgical arm is modeled, the body surface model is traversed, the surgical incision point is calculated, and the surgical path simulation is utilized to perform spatial analysis and collision detection on each point on the path, thereby generating 2-3 best surgical incision positions, performing path planning, and realizing visualization in a three-dimensional view. The "CT data extraction and three-dimensional modeling", "surgical incision position calculation based on spatial voxel judgment", "surface extraction based on voxels", "spatial collision probability analysis based on organ segmentation", and "cone incision positioning" are key technologies for realizing the method. The method overcomes the problems of current lung minimally invasive surgery hole opening strategies, such as dependence on personal experience of doctors, non-standard strategy, high surgical positioning risk, and the like, provides digital incision positioning and path analysis, and realizes achievement visualization, thereby providing a quantifiable and analyzable surgical incision positioning method for minimally invasive surgery, and significantly improving the scientificity and analyzability of minimally invasive surgery.

[0056] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0057] Embodiment one

[0058] The application discloses a CT scanning data surgical incision positioning method based on spatial voxelization judgment, please see Figure 1 , which comprises the following steps.

[0059] S1: acquiring CT data of a patient and performing preprocessing;

[0060] S2: performing three-dimensional modeling and lesion point marking on the preprocessed CT data;

[0061] S3: performing organ segmentation and body surface contour extraction based on the constructed three-dimensional model;

[0062] S4: performing voxelization on the three-dimensional model after organ segmentation and body surface contour extraction by using a ray polygon intersection test method;

[0063] S5: generating an inverted circular cone with a lesion point center as a vertex based on the obtained three-dimensional voxel model, traversing body surface points covered by the inverted circular cone respectively, taking a line connecting a body surface point and the lesion point center as a surgical path, judging the feasibility of the surgical path according to important organ avoidance and puncture angle, screening out feasible skin points, and then screening out candidate incision positions from all the feasible skin points according to preset conditions;

[0064] S6: obtaining a target incision position according to the three-dimensional voxel model and the candidate incision positions.

[0065] In the specific implementation process, S1 can acquire CT data of a patient by using a medical instrument and perform preprocessing.

[0066] In S4, the ray polygon intersection test method is a method for judging whether two polygons intersect, and the basic principle is that a ray is emitted from a point (usually the center of gravity of the polygon), and then the number of edges of the polygon crossed by the ray is calculated. If the number of edges crossed is odd, the point is inside the polygon; otherwise, the point is outside the polygon. A voxel is a combination of a pixel, a volume and an element, which is equivalent to a pixel in a 3D space.

[0067] S5 is to determine the feasibility of the surgical path according to the important organ avoidance and the entry angle, which is to screen the feasible skin points according to the hard constraint condition, and then to screen the candidate incision position from all the feasible skin points according to the preset condition, which is a further optimization of the screened feasible skin points, and the preset condition is a soft constraint condition.

[0068] In S6, the three-dimensional voxel model obtained in S4 and the incision point position (candidate incision position) calculated in S5 are imported into a three-dimensional visualization system to realize visualization, and finally an accurate and visual ideal surgical incision position is obtained.

[0069] In an embodiment, the preprocessed CT data is subjected to three-dimensional modeling and lesion point marking, which includes:

[0070] The preprocessed CT data is read into a program, and the CT data is converted into HU values, which are used to represent the density of tissues in CT scanning;

[0071] The size of each voxel read in is resampled, and the HU values of the resampled image are interpolated using a trilinear interpolation method;

[0072] The interpolated image is modeled using a marching cubes algorithm, specifically: according to the threshold value required by each organ to be extracted, points greater than or equal to the threshold value required by each organ to be extracted are stored, thereby retaining the extracted organ data and forming a number of triangular faces on the cube, and connecting the triangular faces of each voxel; then draw a polygon, set the polygon color, add the polygon to the three-dimensional coordinates, and complete the creation of the three-dimensional model;

[0073] The coordinates of the lesion points are traversed in the three-dimensional model, and the threshold value corresponding to the coordinates of the lesion points is increased by 100.

[0074] In the specific implementation process, three-dimensional modeling can be achieved through the following process:

[0075] The obtained CT data (DICOM format) is read into a program and imported into a matrix with a size of 512*512, and the CT original data is converted into HU values;

[0076] The size of each voxel read in is resampled to 1 cubic millimeter (because the pixel block size in different DICOM files is inconsistent). In order to improve the effect of three-dimensional modeling, the DICOM data after resampling, i.e. the HU values of the image, are interpolated using a trilinear interpolation method, which extracts the original pixel HU values near each target pixel point, takes the weighted average value of HU to assign to the target pixel, quickly and effectively fills the transformed image data, making the image smooth and continuous.

[0077] The mobile cube algorithm is used for modeling. The general principle is as follows: because the HU values of various organs are different, the threshold value required for extracting each organ is extracted according to the target setting, and the points greater than or equal to the threshold value are stored, so as to retain the extracted organ data and form several triangular faces on the cube (voxel), the triangular faces of each voxel are connected, so as to infinitely approach the surface of the real object. Then, the polygon is drawn, the polygon color is set, the polygon is added to the three-dimensional coordinates, and the creation of the three-dimensional model is completed.

[0078] The marking of the lesion point is specifically implemented in the following manner: the marking coordinate point is traversed, the corresponding threshold value is increased by 100, and the deepening marking in the three-dimensional model is realized.

[0079] In an embodiment, the manner of converting the CT data into HU values is as follows:

[0080] HU = val x slope + intercept

[0081] val represents the original pixel value in the CT image, slope is the slope parameter of the CT scanner, and intercept represents the intercept parameter of the CT scanner.

[0082] In an embodiment, organ segmentation and body surface contour extraction are performed based on the constructed three-dimensional model, including:

[0083] The CT data is subjected to HU value conversion and resampling processing; the HU values are mapped to the interval of 0-255, and then the maximum inter-class variance method is used to determine the threshold value; based on the determined threshold value, the contour is found, specifically: when there are both value and 0 points around a point, the point is saved and saved in a newly created array; the newly created array is subjected to 3D modeling, and the contour extraction is realized;

[0084] The threshold value required for separating the lung is set to -600, the image is binarized, and each region is marked; according to the marking, the number of pixels contained in each connected region is counted, two largest connected regions are retained, and then morphological hole filling is performed on the two largest connected regions, so as to obtain the lung model; then, a suitable seed point is selected, and the region growing method is used to obtain the three-dimensional models of the lung organ, the cardiovascular system and the heart, so as to complete organ segmentation.

[0085] In the implementation process, the principle of using the maximum inter-class variance method to determine the threshold is to enumerate the threshold from 0 to 255, divide the image value greater than or less than the threshold into two classes, and calculate the variance. When the variance is the largest, it means that the image can be best divided into two classes, and the value is used as the threshold. Values less than the threshold are changed to 0, and values greater than the threshold are changed to the threshold, to achieve binarization. Then find the contour on this basis. The principle is that if the pixel points around a point have values or are all 0, it means that the point is inside or outside. Therefore, when there are value and 0 points around a point, it means that the point is on the contour. Save the contour points and save them in a new array. Finally, perform 3D modeling on the new array to achieve contour extraction.

[0086] According to the known CT image processing experience value, the HU value of the lung tissue CT is above -600, so the threshold required for separating the lung is set to -600. At this time, the extracted part contains the entire lung, and organ segmentation is required.

[0087] Region growing is an image segmentation method based on pixel similarity. It starts from some seed points and gradually adds similar pixel points to the same region.

[0088] In one embodiment, ray polygon intersection testing is used to voxelize the three-dimensional model after organ segmentation and body surface contour extraction, including:

[0089] A three-dimensional space composed of a plurality of voxels is established;

[0090] For the surface of the three-dimensional model, a projection to two-dimensional method is used to find the maximum and minimum values of the triangular network model in the x, y, and z directions, respectively. Then, a cube with edges parallel to the coordinate axes is established with the maximum and minimum values as the opposite vertices, as the AABB bounding box of the triangular network model. The intersection operation is performed on the AABB bounding box and the triangular network model to obtain the voxel unit intersected by the triangular surface, i.e. the surface of the three-dimensional model.

[0091] For the interior of the three-dimensional model, all empty voxels in the AABB bounding box are traversed, and two rays are emitted from the center position in the axis-aligned direction. The intersection positions of the two rays with the triangular surface are calculated, and the normal vectors of the intersection points are obtained. Then, according to the relationship between the normal vectors of the intersection points, it is determined whether the voxel is located outside or inside the three-dimensional model.

[0092] In the implementation process, the voxelization is roughly divided into two steps of surface voxelization and internal voxelization. Two rays are emitted from the center position of the hollow sphere in the axial direction, and the directions of the two rays are opposite, but the basic directions are axial. The intersection positions of the two rays with the triangular faces are calculated, and the normal vectors of the intersection points are obtained. Then, according to the relationship between the two normal vectors, it is judged that if the directions of the two normal vectors are the same, it is proved that one ray intersects with the triangular face, which indicates that the voxel is located outside the three-dimensional model, otherwise, it indicates that it is inside the three-dimensional model.

[0093] In an embodiment, the preset conditions include a distance between the surgical path extension line and the important organ, a puncture angle of the surgical needle with the skin surface, and a length of a puncture path in the lung.

[0094] Specifically, the center voxel of the lesion point is calculated, a reverse cone with a cone angle of 60° is set with the point as a vertex, a starting direction is set, all skin points covered by the reverse cone are obtained, the skin points are connected to the center of the lesion point, and the feasible skin points are screened out through hard constraint conditions such as important organ avoidance and puncture angle. The puncture angle of the surgical needle with the skin surface and the length of the puncture path in the lung are taken as soft constraint conditions to calculate the weight value of the feasible skin points. In this way, all skin points are traversed, and the three incision points with the optimal comprehensive conditions are selected.

[0095] The specific implementation method of the puncture angle constraint is as follows:

[0096] In the actual clinical operation of penetrating the skin surface, if the puncture angle is too small, that is, the puncture path is too inclined, the surgical needle will be lowered due to the gravity of the handle, which will cause the surgical needle to easily slide on the skin surface, make the path deviate or not easy to puncture, and thus reduce the first puncture success rate. In order to improve the success rate of the operation, the puncture angle of the surgical needle with the skin surface cannot be too small. According to the clinical experience of doctors, the puncture angle needs to be greater than 20° to meet the constraint condition. The vector from the lesion point to the incision position is calculated, the puncture angle is calculated, and the incision position with a puncture angle greater than 20° is deleted.

[0097] In an embodiment, candidate incision positions are screened out from all feasible skin points according to preset conditions, including:

[0098] The collision probability of the surgical motion simulation is calculated according to the distance between the center of the voxelized model of the surgical needle and the important organ, and a collision risk sub-objective function is set;

[0099] A puncture angle risk sub-objective function is set according to the skin puncture angle and the lung puncture angle;

[0100] Set the path risk sub-objective function under the condition of the total length of the puncture path and the length in the lung;

[0101] Set the weight of each risk sub-objective function to obtain the total risk function of the incision position;

[0102] According to the total risk function of the incision position, the weight of each skin point is calculated, and the three positions with the smallest risk are selected as the candidate incision positions.

[0103] The specific implementation method of checking the collision interference in the working space of the surgical needle is as follows:

[0104] The surgical needle is abstracted into a three-dimensional conical object, and the center of the voxel block is calculated. The distance between the center of the surgical needle and the centers of the extracted organs is calculated. If the distance is less than the length of the surgical needle, it is considered that there is a high probability of collision between the surgical needle and the organs during the puncture process, and the incision position is discarded. Otherwise, it is considered that the incision position has a certain feasibility. According to the distance, a risk sub-objective function D(i) is established using the formula:

[0105]

[0106] Where D i is the distance between the current path (path i) and the important organs, D min is the minimum value of the distance between all effective paths and important organs, and D max is the maximum value of the distance between all effective paths and important organs. According to the definition of the risk sub-objective function D(i), the greater the distance between the path and the important organs, the smaller the value of the sub-objective function, and the smaller the risk of the corresponding path.

[0107] In the process of using hard constraint conditions to screen the incision, in order to avoid the slipping of the surgical needle, the paths with too small skin and lung parenchyma surface insertion angles are screened out through the insertion angle constraint. For paths that meet the constraint conditions, vertical skin insertion of the needle, that is, when the insertion angle of the surgical needle and the skin surface is 90°, the surgical needle is most easy to grasp, can better maintain the angle of puncture, and is also convenient for material taking. Relevant literature shows that in the actual operation of lung puncture surgery, vertical angle insertion compared with oblique angle significantly improves the success rate of obtaining lesion tissue in one puncture, and the statistical analysis has significant difference (P<0.05), and reduces the complications, that is, the greater the insertion angle of the surgical needle and the skin and lung parenchyma, the closer to 90°, the lower the risk of surgery, and the better the incision position. Therefore, the skin insertion angle and the lung parenchyma insertion angle are both soft constraint conditions for risk assessment of the incision position. The risk sub-objective functions under the conditions of the skin insertion angle (Insert Angle of Skin, IAS) and the lung parenchyma insertion angle (Insert Angle of Lung, IAL) are defined, and the calculation formula is as follows:

[0108]

[0109] wherein, a i , b i are the angles of entry into the skin and the parenchyma of the lung for the current path (path i), a max , b max are the minimum values of the angles of entry into the skin and the parenchyma of the lung for all valid paths, a min , b min are the maximum values of the angles of entry into the skin and the parenchyma of the lung for all valid paths. From the definition of the risk sub-objective function, the greater the angles of entry into the skin and the parenchyma of the lung, the smaller the value of the sub-objective function, and the lower the risk of surgery for the incision.

[0110] The shorter the length of the puncture path, the smaller the damage to the patient and the lower the risk of the path. For lung puncture surgery, the total length of the path (the length from the entry point on the skin to the target point of the tumor) and the length of the path in the lung (the length from the entry point on the surface of the parenchyma of the lung to the target point of the tumor) are two length conditions that are concerned by the clinician in clinical operation. The total length and the length in the lung of each path need to be calculated, and the sub-objective function of the risk of the path under these two target conditions is defined as two sub-objectives of the optimization of the incision position. The intersection of path i and the surface of the parenchyma of the lung is Q i (a i , b i , c i ), the corresponding entry point on the skin is P i (x i , y i , z i ), and the target point of the lesion is O(a0, b0, c0). The total length of the puncture path (Trajectory Length, TL) is:

[0111]

[0112] The length of the puncture path in the lung (Trajectory Length in Lung, TLL) is:

[0113]

[0114] The two lengths are normalized using the method of standard deviation, and the sub-objective function of the risk of the path under the conditions of the total length of the puncture path and the length in the lung is defined. The calculation formula is as follows:

[0115]

[0116] wherein, L i , L2i is the total length of the current path and the intrapulmonary length, L min , L 2min is the minimum value of the total length of all effective paths and the intrapulmonary length, L max , L 2max is the maximum value of the total length of all effective paths and the intrapulmonary length. As can be known from the definition of the risk sub-objective function, the greater the total length of the path and the intrapulmonary length, the greater the value of the sub-objective function, and the higher the risk of the path.

[0117] Finally, the total risk function of the incision position is obtained, and the calculation formula is as follows:

[0118]

[0119] Where w is the weight, which is set according to the actual operation needs, and initially, the weights of the elements are equal, and the values are all 1.0. R is the above five risk sub-objective functions. Finally, the weight of each skin point is calculated, and the three positions with the minimum risk are selected as the candidate incision positions.

[0120] The automatic selection of the incision position is visualized, scientific and accurate, convenient and fast, greatly reduces the operation difficulty and operation risk, improves the operation quality and safety, and greatly relieves the patient's pain.

[0121] The specific implementation steps of the present application will be described below through specific examples:

[0122] The technical scheme provided by the present application is a method based on CT scan data, using a moving cube algorithm for three-dimensional modeling and a test method of ray and polygon intersection to voxelize the three-dimensional model, and obtaining the best incision position through feasibility analysis and weight addition, which can be used for various scenes of lung minimally invasive incision surgery. Among them, "extraction of CT data and three-dimensional modeling" and "surgical incision position calculation based on spatial voxel judgment" are the key technologies for realizing the method. The technical process of the present application will be described in detail below.

[0123] Step 1: Obtain the CT data of the patient using a medical instrument and perform preprocessing;

[0124] Read the DICOM file using the PyDICOM library (an open source Python library dedicated to reading, writing and manipulating Digital Imaging and Communications in Medicine (DICOM) files), and then sort the z-axis coordinates of the slices in the patient coordinate system as shown in FIG. 1. Figure 2 Then calculate the z-axis position of the adjacent slices to calculate the thickness of each slice. The z-axis coordinate of the current slice minus the z-axis coordinate of the last slice is the thickness of the current slice.

[0125] If the data is lost, it is replaced by the coordinates of the slice relative to the scanner. The thickness of the current slice is the coordinates of the current slice relative to the scanner minus the coordinates of the previous slice relative to the scanner.

[0126] Step 2: three-dimensional modeling and marking of lesion points are performed on the data obtained in step 1; the CT slice data obtained in step 1 is stacked along the z axis to form a three-dimensional array, thereby forming a complete CT image data set. Subsequently, HU value conversion is performed.

[0127] Because the pixel block sizes in different DICOM files are inconsistent, the size of each read-in voxel is resampled and converted to 1 cubic millimeter.

[0128] First, the pixel spacing (Pixel Spacing) in the scanning data is obtained, which is usually a two-dimensional list containing the pixel spacings in the x and y directions. Then, the slice thickness calculated in step 1 is added, and the list is converted into an array for subsequent calculation. Assuming that the converted array is named spacing, the scaling factor resize_factor = spacing / new_spacing (the desired pixel block size, which is [1, 1, 1] in this case) is calculated. Because the number of pixels must be an integer during operation, the result may be a floating-point number after multiplying the original array size by the scaling factor (resize_factor). Therefore, the size of the new image is calculated by multiplying the original image size by the scaling factor, and the image size is obtained by taking the integer part to make the number of pixels an integer. Then, the new scaling factor is calculated by dividing the rounded image size by the original image size, so that the scaling factor that makes the number of pixels after scaling an integer is obtained.

[0129] In order to improve the three-dimensional modeling effect, the data is resampled and interpolated, and the trilinear interpolation method is used here. The basic principle is to extract the original pixel data near each target pixel point, take the average value, and assign it to the target pixel. Let the target pixel point be (i, j, k), find the nearest 8 original pixel points, calculate the weight according to the distance, and finally assign the value to the target pixel. That is where c(i0, j0, k0) represents the HU value of the target point, c(i, j, k) represents the HU value of one of the 8 original pixel points near the target point, P n and represents the corresponding weight. In this way, the transformed image data is filled quickly and effectively, making the image smooth and continuous.

[0130] Then, the marching cubes algorithm is used for modeling. The basic principle is as follows: according to the target setting threshold, the points greater than or equal to the threshold are stored, thereby forming a cube (voxel) as shown inFigure 3 The several triangular faces shown respectively store the vertex coordinates of the triangular faces and the face index. Then the triangular faces of each voxel are connected to approximate the surface of the real object. Finally, the drawing of the triangular faces is added to the three-dimensional coordinates to complete the creation of the three-dimensional model.

[0131] Then the lesion coordinate points are traversed, and the corresponding threshold is increased by 100 to achieve deep marking in the three-dimensional model.

[0132] Step 3: On the basis of step 2, organ segmentation and skin contour extraction are performed;

[0133] First, the obtained DICOM data is subjected to the above HU value conversion and resampling processing. Then the HU value is mapped to the interval of 0-255, i.e. (value-min) / (max-min)*255. Subsequently, the OTSU algorithm is used to determine the threshold. The general principle is to enumerate the threshold from 0-255, divide the image value into two categories and calculate the variance, and when the variance is the largest, the image can be best divided into two categories, and the value is used as the threshold. Values less than the threshold are changed to 0, and all values greater than the threshold are changed to the threshold, achieving binarization. Then find the contour on this basis. The general principle of judgment is that if the pixel points around a point all have values or are all 0, it means that the point is inside or outside. Therefore, when there are value and 0 points around a point, it means that the point is on the contour. Save the contour points and save them in a newly created array. Finally, perform 3d modeling on the newly created array (method same as above) to achieve contour extraction. The effect is shown in Figure 4

[0134] According to the feature that the HU values of different tissues in the human body are different, the threshold is set to -600. After binarizing the image, all pixel points are traversed based on depth-first search and each connected region is labeled. According to the label, the number of pixels contained in each connected region is counted, and the remaining small noise points are ignored, and the two largest connected regions are retained. Then perform morphological closing operation on the image to obtain the lung model. The specific principle is to first perform dilation operation on the image and then perform erosion operation, as shown in Figure 5 ​The principle of expansion operation is to traverse the image with a kernel (square matrix) one by one, if there is a point with value not equal to 0 on the image intersecting with the point with matrix element equal to 1, then the central element value of the image is changed to 1, realizing the filling of the hole and the expansion of the image. The principle of erosion operation is similar, traverse the image pixel points with the same kernel, if there is a point with value equal to 0 on the image intersecting with the point with matrix element equal to 0, then the central element of the image is changed to 0, so that the internally filled holes are retained because of the values on the periphery, and the newly added pixel points on the edge are removed by the erosion operation, thus realizing the filling of the hole. Then select a representative pixel point in the image as a suitable seed point, in this example, the center point is selected, and the region growing method is used, that is, starting from the seed point, traverse the remaining pixel points on the image, judge the similarity with the seed point, in this example, the HU value is used as the judgment basis, if the difference between the HU value of the pixel point and the HU value of the seed point is less than the threshold, it is considered as the same region, and is added to the set, finally the three-dimensional model of the lung organ, the cardiovascular system and the heart is obtained, and the organ segmentation is completed.

[0135] Step 4: Use the three-dimensional model obtained in step 3, as shown in Figure 6 , and use the ray polygon intersection test method for voxelization. Voxelization is roughly divided into two steps of surface voxelization and internal voxelization. First, a three-dimensional space composed of a plurality of voxels is established. For the surface of the three-dimensional model, a projection to two-dimensional method is used to find the maximum and minimum values of the triangular network model in x, y and z directions respectively, and then a cube with edges parallel to the coordinate axes is established with the maximum and minimum values as the opposite vertices, that is, the AABB bounding box of the triangular network model. Then, the intersection operation is performed on the AABB and the triangular network model to obtain the voxel unit intersected by the triangular face, that is, the surface of the three-dimensional model. For the interior of the three-dimensional model, traverse all the empty voxels in the AABB model, and emit two rays from the center position in the axis-aligned direction. The directions of the two rays are opposite, but the basic directions are axis-aligned. Calculate the intersection positions of the two rays with the triangular face, and obtain the normal vectors of the intersection points. Then, according to the relationship between the two normal vectors, it is judged that if the directions of the two normal vectors are the same, it is proved that one ray intersects with the triangular face, which means that the voxel is located outside the three-dimensional model, otherwise, it means that it is inside the three-dimensional model. The effect is shown in Figure 7 .

[0136] Step 5: As shown in Figure 8 , generate an inverted cone with the center of the lesion point as the vertex on the basis of the voxel model, traverse the skin points covered by the cone respectively, and for each skin point, draw a line from the center of the lesion point, use the hard constraint condition to judge the feasibility of the surgical path, then add the weights of the voxel blocks on the line and use the collision probability detection of motion simulation under the surgical condition to calculate the distance between the voxelized model center of the surgical needle and the important organs, and the entry angle (such as Figure 8The three surgical incision points are obtained and stored as a group according to the soft constraint conditions such as the length of the puncture path, and then the best three points (candidate incision positions) are screened out after calculating the risk coefficients of all skin points.

[0137] Step 6: The three-dimensional model obtained in step 4 and the surgical incision point calculated in step 5 are imported into a three-dimensional application system to realize visualization, and finally the ideal surgical incision position is obtained.

[0138] Embodiment Two

[0139] Based on the same inventive concept, the embodiment discloses a CT scanning data surgical incision positioning device based on spatial voxelization judgment, which comprises:

[0140] A data acquisition module is configured to acquire CT data of a patient and perform preprocessing.

[0141] A three-dimensional modeling module is configured to perform three-dimensional modeling on the preprocessed CT data and mark a lesion point.

[0142] A segmentation and extraction module is configured to perform organ segmentation and body surface contour extraction based on the constructed three-dimensional model.

[0143] A voxelization module is configured to perform voxelization on the three-dimensional model after organ segmentation and body surface contour extraction by using a ray polygon intersection test method.

[0144] A candidate incision position screening module is configured to generate an inverted circular cone with the center of the lesion point as the vertex based on the obtained three-dimensional voxel model, traverse the body surface points covered by the cone respectively, take the line connecting the body surface point and the center of the lesion point as a surgical path, judge the feasibility of the surgical path according to important organ avoidance and puncture angle, screen out feasible skin points, and then screen out candidate incision positions from all feasible skin points according to preset conditions.

[0145] A surgical incision positioning module is configured to obtain a target incision position according to the three-dimensional voxel model and the candidate incision position.

[0146] Since the device introduced in the embodiment two is used to implement the CT scanning data surgical incision positioning method based on spatial voxelization judgment in the embodiment one, the specific structure and deformation of the device can be understood by those skilled in the art based on the method introduced in the embodiment one, and thus will not be described here again. Any device used in the method in the embodiment one belongs to the scope of the present application.

[0147] Embodiment Three

[0148] Based on the same inventive concept, the application further provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to implement the method in embodiment one.

[0149] Since the computer readable storage medium introduced in embodiment three of the application is the computer readable storage medium used for implementing the CT scan data surgical incision positioning method based on spatial voxelization judgment in embodiment one of the application, the specific structure and deformation of the computer readable storage medium can be understood by those skilled in the art based on the method introduced in embodiment one of the application, and thus will not be described here again. Any computer readable storage medium used for the method in embodiment one of the application belongs to the scope of the application.

[0150] Embodiment four

[0151] The application further provides a computer device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the method in embodiment one when executing the program.

[0152] Since the computer device introduced in embodiment four of the application is the computer device used for implementing the CT scan data surgical incision positioning method based on spatial voxelization judgment in embodiment one of the application, the specific structure and deformation of the computer device can be understood by those skilled in the art based on the method introduced in embodiment one of the application, and thus will not be described here again. Any computer device used for the method in embodiment one of the application belongs to the scope of the application.

[0153] Those skilled in the art should understand that the embodiments of the application can be provided as a method, a system, or a computer program product. Therefore, the application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the application can adopt a computer program product in the form of being implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program codes.

[0154] The application is described with reference to flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to produce a machine, so that the instructions executed by the computer or other programmable data processing devices produce a device for implementing the functions described in the flowcharts and / or block diagrams. Figure 1 one flow or multiple flows and / or blocksFigure 1 means for performing the function specified in the block or blocks.

[0155] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the preferred embodiments by those of skill in the art once they have the benefit of the present disclosure. Therefore, the appended claims are intended to encompass within their scope all possible variations and modifications of the preferred embodiments. It will be apparent to those of ordinary skill in the art that various changes and modifications can be made to the preferred embodiments without departing from the spirit and scope of the application. Thus, the present application is not intended to be limited to the preferred embodiments but wants to include all such changes and modifications as fall within the scope of the application.

Claims

1. A CT scan data incision positioning method based on spatial voxelization judgment, characterized in that, The application relates to a method for selecting a target incision position based on CT data, comprising the following steps: acquiring CT data of a patient and pre-processing the CT data; performing three-dimensional modeling and lesion point marking on the pre-processed CT data; performing organ segmentation and body surface contour extraction based on the constructed three-dimensional model; performing voxelization on the three-dimensional model after organ segmentation and body surface contour extraction by using a ray polygon intersection test method; generating an inverted circular cone with the center of the lesion point as the vertex based on the obtained three-dimensional voxel model, respectively traversing the body surface points covered by the circular cone, taking the line connecting the body surface point and the center of the lesion point as a surgical path, judging the feasibility of the surgical path according to important organ avoidance and puncture angle, screening out feasible skin points, and then screening out candidate incision positions from all the feasible skin points according to preset conditions; obtaining the target incision position according to the three-dimensional voxel model and the candidate incision position; wherein the preset conditions include the distance between the extension line of the surgical path and the important organ, the puncture angle between the surgical needle and the skin surface, and the length of the lung puncture path; the candidate incision positions are screened out from all the feasible skin points according to the preset conditions, which comprises the following steps: calculating the collision probability of the surgical motion simulation according to the distance between the center of the surgical needle voxelization model and the important organ, and setting a collision risk sub-objective function, which is specifically as follows: wherein is the distance of the current path to the vital organ, is the minimum of the distance of all valid paths to the vital organ, is the maximum of the distance of all valid paths to the vital organ, the current path being path i; setting a puncture angle risk sub-objective function according to the skin puncture angle and the lung puncture angle, which is specifically as follows: wherein, , are the current path and the skin's angle of entry, the lung's parenchymal angle of entry, respectively, are the minimum of all valid path and skin's angle of entry, the minimum of the lung's parenchymal angle of entry, are the maximum of all valid path and skin's angle of entry, the maximum of the lung's parenchymal angle of entry; setting a path risk sub-objective function under the conditions of the total length of the puncture path and the length in the lung, which is specifically as follows: wherein, , is the total length of the current path and the intrapulmonary length, , is the minimum of the total length of all valid paths and the intrapulmonary length, , is the maximum of the total length of all valid paths and the intrapulmonary length; setting the weight of each risk sub-objective function to obtain a total risk function of the incision position, which is specifically as follows: wherein is a weight for the nth risk sub-objective function, is the nth risk sub-objective function; calculating the weight of each skin point according to the total risk function of the incision position, and selecting three positions with the smallest risk as the candidate incision positions.

2. The CT scan data incision positioning method based on spatial voxelization judgment of claim 1, wherein, The three-dimensional modeling and lesion point marking on the pre-processed CT data comprises the following steps: reading the pre-processed CT data into a program, and converting the CT data into HU values, which are used to represent the density of tissues in CT scanning; resampling the size of each voxel, and using a trilinear interpolation method to perform interpolation processing on the HU values of the resampled image; modeling the image after the interpolation processing by using a moving cube algorithm, which is specifically as follows: setting the threshold value required by each organ to be extracted according to the target, storing the points greater than or equal to the threshold value required by each organ to be extracted, thereby retaining the extracted organ data and forming a plurality of triangular faces on the cube, connecting the triangular faces of each voxel; then drawing a polygon, setting the color of the polygon, adding the polygon to the three-dimensional coordinates, and completing the creation of the three-dimensional model; traversing the coordinate points of the lesion point in the three-dimensional model, and increasing the HU value threshold corresponding to the coordinate points of the lesion point by 100.

3. The CT scan data incision positioning method based on spatial voxelization judgment of claim 2, wherein, The way of converting the CT data into HU values is as follows: represents an original pixel value in a CT image, is a slope parameter of a CT scanner, is an intercept parameter of a CT scanner.

4. The CT scan data incision positioning method based on spatial voxelization judgment of claim 1, wherein, The organ segmentation and body surface contour extraction based on the constructed three-dimensional model comprises the following steps: performing HU value conversion and resampling processing on the CT data; mapping the HU values to the interval of 0-255, and then determining the threshold value by using the maximum inter-class variance method; finding the contour based on the determined threshold value, which is specifically as follows: when a point has both value and 0 points around it, the point is saved and stored in a newly created array; performing 3D modeling on the newly created array to realize contour extraction. The threshold value required for separating the lung is set as -600, the image is binarized, each region is labeled, the number of pixels contained in each connected region is counted according to the label, two largest connected regions are reserved, and then the two largest connected regions are filled with holes by using morphological operation to obtain a lung model; then, a suitable seed point is selected, and a region growing method is used to obtain three-dimensional models of the lung organ, cardiovascular system and heart to complete organ segmentation.

5. The CT scan data incision positioning method based on spatial voxelization judgment of claim 1, wherein, The three-dimensional model after organ segmentation and body surface contour extraction is voxelized using a ray polygon intersection test method, including: A three-dimensional space composed of a plurality of voxels is established; For the surface of the three-dimensional model, the maximum and minimum values of the triangular network model in the x, y and z directions are found by projecting onto the two dimensions, respectively, and then a cube with each side parallel to the coordinate axis is established with the maximum and minimum values as the opposite vertices, as the AABB bounding box of the triangular network model; the intersection operation is performed on the AABB bounding box and the triangular network model to obtain the voxel unit intersected by the triangular surface, i.e. the surface of the three-dimensional model; For the interior of the three-dimensional model, all empty voxels in the AABB bounding box are traversed, and two rays are emitted from the center position in the axis-aligned direction, the intersection positions of the two rays with the triangular surface are calculated, and the normal vectors of the intersection points are obtained; then, it is judged whether the voxel is located outside or inside the three-dimensional model according to the relationship between the normal vectors of the intersection points.

6. A CT scan data incision positioning device based on spatial voxelization judgment, characterized in that, It includes: A data acquisition module for acquiring CT data of a patient and pre-processing the CT data; A three-dimensional modeling module for three-dimensional modeling of the pre-processed CT data and lesion point labeling; A segmentation and extraction module for organ segmentation and body surface contour extraction based on the constructed three-dimensional model; A voxelization module for voxelizing the three-dimensional model after organ segmentation and body surface contour extraction using a ray polygon intersection test method; A candidate incision position screening module for generating an inverted cone with the center of the lesion point as the vertex based on the obtained three-dimensional voxel model, traversing the body surface points covered by the cone respectively, taking the line connecting the body surface point and the center of the lesion point as a surgical path, judging the feasibility of the surgical path according to important organ avoidance and puncture angle, screening out feasible skin points, and then screening out candidate incision positions from all feasible skin points according to preset conditions; A surgical incision positioning module for obtaining a target incision position according to the three-dimensional voxel model and the candidate incision position; The preset conditions include the distance between the surgical path extension line and the important organ, the puncture angle of the surgical needle with the skin surface, and the length of the lung puncture path; the candidate incision positions are screened out from all feasible skin points according to the preset conditions, including: The collision probability of the surgical motion simulation is calculated according to the distance between the center of the skin voxelization model and the important organ, and a collision risk sub-objective function is set, which is specifically: wherein is the distance of the current path to the vital organ, is the minimum of the distance of all valid paths to the vital organ, is the maximum of the distance of all valid paths to the vital organ, the current path being path i; The puncture angle risk sub-objective function is set according to the skin puncture angle and the lung puncture angle, which is specifically: wherein, , are the current path and the skin's angle of entry, the lung's parenchymal angle of entry, respectively, are the minimum of all valid path and skin's angle of entry, the minimum of the lung's parenchymal angle of entry, are the maximum of all valid path and skin's angle of entry, the maximum of the lung's parenchymal angle of entry; The path risk sub-objective function is set under the conditions of the total length of the puncture path and the length in the lung, which is specifically: wherein, , is the total length of the current path and the intrapulmonary length, , is the minimum of the total length of all valid paths and the intrapulmonary length, , is the maximum of the total length of all valid paths and the intrapulmonary length; The weight of each risk sub-objective function is set to obtain a total risk function of the incision position, which is specifically: wherein is a weight for the nth risk sub-objective function, is the nth risk sub-objective function; The weight of each skin point is calculated according to the total risk function of the incision position, and three positions with the minimum risk are selected as candidate incision positions.

7. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the CT scan data surgical incision positioning method based on spatial voxelization judgment as claimed in any one of claims 1 to 5.

8. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the CT scan data surgical incision positioning method based on spatial voxelization judgment as claimed in any one of claims 1 to 5.

Images