A registration accuracy verification method and apparatus

Abstract

This application discloses a registration accuracy verification method and apparatus, relating to the field of data processing technology. It acquires three-dimensional tomographic images of a surgical area of ​​a bone model using a first image acquisition device and registration and marking components. A reference relationship matrix between the coordinate systems is calculated based on the first coordinate point set of the first marker of the registration component in the view coordinate system and the second coordinate point set in the world coordinate system. Based on this matrix, the mapped coordinate point set of the fourth coordinate point set of the second marker of the marking component in the world coordinate system is calculated and then mapped to the view coordinate system. The registration accuracy of the three-dimensional tomographic images acquired by the first image acquisition device is verified based on the distance error, angle error, and planar error between the third coordinate point set of the second marker of the marking component in the view coordinate system and the mapped coordinate point set. This application achieves objective and accurate verification of registration accuracy independent of the surgical navigation system, improving the reliability and versatility of registration accuracy verification.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and more specifically, to a registration accuracy verification method and apparatus. Background Technology

[0002] Surgical navigation and positioning systems spatially register preoperative medical images, such as computed tomography (CT) and magnetic resonance imaging (MRI), with the patient's actual anatomical structures during surgery. This provides surgeons with a fluoroscopic visualization view during the operation and is a key technology for improving surgical accuracy and safety. The core of the spatial registration algorithm is to unify the virtual space (image coordinate system) with the physical space (actual coordinate system).

[0003] Existing registration algorithms, such as the Iterative Closest Point (ICP) algorithm based on surface point clouds and feature point-based registration algorithms, typically output a registration error value, such as the root mean square error (RMSE), after calculating the spatial transformation matrix. However, this RMSE is a theoretical fitting residual calculated by the algorithm based on its matched input point set. It is calculated by the navigation system itself, and its evaluation process is highly coupled with the registration calculation process, lacking objectivity and failing to directly reflect the position and angular deviation of surgical instruments relative to the patient's anatomical structures in real physical space. Furthermore, it cannot effectively assess the extrapolation accuracy of the algorithm in areas not involved in the calculation (especially clinically critical operative areas). This leads to a significant deviation between the theoretically high-precision registration and the clinically most important surgical target area. Therefore, there is an urgent need for an objective physical verification method independent of the navigation system itself to solve these technical problems. Summary of the Invention

[0004] In view of this, this application proposes a registration accuracy verification method and apparatus. By non-uniformly embedding verification points in the bone model, the registration point set and marker point set in the view coordinate system and the world coordinate system are obtained respectively. The reference relationship matrix of the two coordinate systems is calculated, thereby obtaining the marker point set and the mapped coordinate point set in the view coordinate system. By calculating the distance error, angle error and plane error between the two, the registration accuracy is verified. This achieves objective and accurate verification of registration accuracy without relying on the surgical navigation system, and improves the reliability and versatility of registration accuracy verification in surgical navigation technology.

[0005] Firstly, this application proposes a registration accuracy verification method, comprising the following steps: Acquire three-dimensional tomographic images of the surgical area of ​​the bone model using the registration and marking components of the first image acquisition device; Acquire three-dimensional tomographic images of the surgical area of ​​the bone model using the registration and marking components of the first image acquisition device; The reference relationship matrix between the coordinate systems is calculated based on the first set of coordinate points of the first marker in the view coordinate system and the second set of coordinate points in the world coordinate system of the registration component. The mapped coordinate point set of the fourth coordinate point set in the view coordinate system is calculated based on the fourth coordinate point set of the second marker of the marker component in the world coordinate system and the reference relationship matrix between the coordinate systems; Calculate the distance error, angle error, and planar error between the third set of coordinate points of the second marker of the marking component in the view coordinate system and the mapped coordinate point set, wherein the angle error is the difference in the angle between the direction vectors of at least two points in the third set of coordinate points and the mapped coordinate point set before and after registration; the planar error is the difference in the plane normal vectors between three non-collinear points in the third set of coordinate points and the mapped coordinate point set. The accuracy of the registration of the three-dimensional tomographic images acquired by the first image acquisition device is verified based on the distance error, angle error, and planar error.

[0006] Optionally, the angle error is the difference in angle between the direction vectors of the third coordinate point set and at least two points in the mapped coordinate point set before and after registration, including: The angle difference between the direction vectors before and after registration can be calculated by selecting the first and second marker points in the third coordinate point set. Alternatively, the angle difference between the direction vectors before and after registration can be calculated by selecting at least two marker points with axial distribution from the third coordinate point set.

[0007] Optionally, the angle error is the difference in angle between the direction vectors of the third coordinate point set and at least two points in the mapped coordinate point set before and after registration, including: The angle difference between the direction vectors before and after registration can also be calculated by selecting at least two pairs of marker points from the third coordinate point set.

[0008] Optionally, the angle difference between the direction vectors before and after registration can also be calculated by selecting at least two pairs of marker points from the third coordinate point set, including: Calculate the angle difference between at least two pairs of marked points, and take the average or maximum value as the final evaluation index.

[0009] Optionally, the plane error is the difference in plane normal vectors between the third set of coordinate points and three non-collinear points in the mapped set of coordinate points, including: Based on the reference plane formed by the three non-collinear points in the third coordinate point set, the normal angle and tilt direction of the reference plane before and after registration are calculated and visualized.

[0010] Optionally, the calculation and visualization of the normal angle and tilt direction of the reference plane before and after registration includes: Based on the error information obtained from the visualization, an optimization strategy for the error information is generated. The optimization strategy is used to adjust the execution mode of the subsequent registration method.

[0011] Optionally, the second markers of the marking component are arranged in a locally dense, non-uniform distribution, wherein the surgical area and the edge away from the registration component have a higher marker density than other areas.

[0012] Secondly, this application also proposes a registration accuracy verification device, comprising: The image acquisition unit is used to acquire three-dimensional tomographic images of the surgical area of ​​the bone model acquired by the registration and marking components of the first image acquisition device. The coordinate transformation unit is used to calculate the reference relationship matrix between the coordinate systems based on the first set of coordinate points of the first marker of the registration component in the view coordinate system and the second set of coordinate points in the world coordinate system. The coordinate acquisition unit is used to calculate the mapped coordinate point set of the fourth coordinate point set in the view coordinate system based on the fourth coordinate point set of the second marker of the marker component in the world coordinate system and the reference relationship matrix between the coordinate systems; The error calculation unit is used to calculate the distance error, angle error, and planar error between the third set of coordinate points of the second marker of the marking component in the view coordinate system and the mapped coordinate point set, wherein the angle error is the difference in the angle between the direction vectors of at least two points in the third set of coordinate points and the mapped coordinate point set before and after registration; the planar error is the difference in the plane normal vectors between three non-collinear points in the third set of coordinate points and the mapped coordinate point set. The verification unit is used to verify the accuracy of the registration of the three-dimensional tomographic images acquired by the first image acquisition device based on the distance error, angle error, and plane error.

[0013] This application can achieve at least the following beneficial effects: This application discloses a registration accuracy verification method and apparatus. By non-uniformly embedding verification points in a bone model, it obtains registration point sets and marker point sets in the view coordinate system and the world coordinate system, respectively. It then calculates the reference relationship matrix between the two coordinate systems to obtain the marker point set and the mapped coordinate point set in the view coordinate system. By calculating the distance error, angle error, and plane error between the two, it verifies the registration accuracy. This method achieves objective and accurate verification of registration accuracy without relying on the surgical navigation system, thus improving the reliability and versatility of registration accuracy verification in surgical navigation technology. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as limiting the scope of this application.

[0015] Figure 1 This is a bone model and a three-dimensional tomographic image of a registration accuracy verification method according to an embodiment of this application; Figure 2 This is a schematic flowchart of a registration accuracy verification method according to an embodiment of this application; Figure 3 This is a partial flowchart of a registration accuracy verification method according to an embodiment of this application; Figure 4 This is another three-dimensional tomographic image of a registration accuracy verification method according to an embodiment of this application; Figure 5 This is another three-dimensional tomographic image of a registration accuracy verification method according to an embodiment of this application; Figure 6 This is a schematic diagram of the registration accuracy verification device according to an embodiment of this application; Figure 7 This is a partial structural schematic diagram of a registration accuracy verification device according to another embodiment of this application.

[0016] Reference numerals: 10 - bone model; 110 - first marker; 120 - second marker. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. However, it should be understood that the described embodiments are merely some exemplary embodiments of this application, and not all embodiments. Therefore, the following detailed description of the embodiments of this application is not intended to limit the scope of protection claimed by this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0018] As mentioned above, existing registration algorithms lack objective, high-precision, and targeted physical verification methods and cannot assess the true pose error of the surgical area. Therefore, this application proposes a registration accuracy verification method and apparatus. By non-uniformly embedding verification points in a bone model, registration point sets and marker point sets in the view coordinate system and the world coordinate system are obtained respectively. The reference relationship matrix between the two coordinate systems is calculated, thereby obtaining the marker point set and the mapped coordinate point set in the view coordinate system. By calculating the distance error, angle error, and plane error between the two, the registration accuracy is verified. This achieves objective and accurate verification of the registration accuracy of the surgical navigation system, improving the reliability and versatility of registration accuracy verification in surgical navigation technology.

[0019] Figure 1 This refers to a bone model and its three-dimensional tomographic image to which the registration accuracy verification method and apparatus of this application are applicable. Figure 2 This is a schematic flowchart of a registration accuracy verification method according to an embodiment of this application, as shown below. Figure 2 As shown, the method may include the following steps: Step S201: Obtain three-dimensional tomographic images of the surgical area of ​​the bone model acquired by the first image acquisition device using the registration and marking components.

[0020] In this embodiment, anterior cruciate ligament reconstruction (ACL) surgery is used as an example to illustrate the preparation of a dedicated verification bone model and the strategy for pre-embedding verification points. The first image acquisition device can be a CT scanner, MRI scanner, cone-beam computed tomography (CBCT) device, etc. A polymer material (such as polyurethane foam or special 3D printing resin) with biomechanical properties similar to human bones is selected. A tibial model consistent with the anatomical morphology of the upper end of the human tibia is prepared by high-precision CNC machining or 3D printing technology based on patient CT data. This model needs to accurately reproduce the articular surface morphology of the tibial plateau, the intercondylar ridge, and the predetermined position of the tibial bone tunnel of the ACL reconstruction to ensure that the verification environment has a high degree of clinical reproducibility. High-density radiopaque spheres with a diameter of 8-12 mm (preferably 10 mm) are selected as verification markers. Verification points are pre-embedded in the target area (both sides of the bone tunnel) of the bone model. The spheres can be made of ceramic, and their grayscale values ​​(HU values) in CT images should show a significant contrast with the surrounding bone model material (usually a difference greater than 1000 HU) to ensure that they can be clearly and accurately identified and segmented by automatic or semi-automatic algorithms in subsequent CT images. In addition to spheres, verification markers can also be set in various shapes such as cylinders, prisms (cubes), frustums, and grooves to adapt to different surgical scenarios. This application does not limit the specific shape of the verification markers.

[0021] Based on the principle of focusing on the target area for evaluation and extrapolating to verify the edge area, a non-uniform, targeted pre-embedding strategy is adopted. Holes matching the diameter of the markers are drilled in the tibial model, the markers are pre-embedded in them, and fixed with biocompatible adhesive. The pre-embedding points are mainly divided into three categories. The density of markers is higher in the surgical area and at the edges far from the registration components than in other areas. Dense pre-embedding is performed in the target surgical area, which is the core clinical operating area, i.e., the area around the pre-designated entrance and exit of the tibial tunnel. Points are densely distributed around the inner and outer sides, as well as the anterior and posterior edges of the tibial tunnel. Preferably, 10 verification points are pre-embedded. This distribution method allows the final verification results to directly and densely reflect the registration accuracy of the actual path (tunnel) of the surgical instruments (such as drills and grafts), avoiding masking errors. Additional pre-embedded points are placed in the model edge region. The model edge region refers to the edge of the model away from the bone tunnel and anatomical verification points commonly used for registration (such as the intercondylar eminence). For example, the anterior, posterior, and distal edges of the tibia model. 4-6 verification points are pre-embedded in this region to evaluate the spatial extrapolation capability and error drift of the registration algorithm in areas far from the registration point set. Appropriate pre-embedded points are placed in the registration area. The registration area is an easily identifiable region on the model surface that is usually selected as a registration point, such as the highest points of the medial and lateral condyles of the tibial plateau and the intercondylar eminence. 1-2 verification points can be pre-embedded in this region. In subsequent verification processes, the coordinates of these pre-embedded points will strictly not participate in the calculation of the registration transformation matrix; they will only be used as independent marker points, i.e., marker components.

[0022] The pre-embedded tibial model is placed in a CT or MRI scanner and scanned using a clinically common orthopedic scanning protocol to obtain a three-dimensional tomographic image containing clear images of markers. The three-dimensional tomographic image data is then imported into three-dimensional reconstruction software (such as Mimics, 3D Slicer), and the surface of the bone model and markers are extracted through threshold segmentation to obtain the registration components and marker components in the view coordinate system.

[0023] This embodiment solves the problem in the prior art where the uniform distribution of verification points leads to the averaging and masking of the accuracy of the target area, enabling the verification results to truly reflect the actual registration accuracy of the surgical operation area. Furthermore, the marker component does not participate in the registration calculation and is only used for verification, ensuring the independence and objectivity of the verification results and avoiding the overfitting problem caused by using the same set of points for self-verification.

[0024] Step S202: Calculate the reference relationship matrix between the coordinate systems based on the first set of coordinate points of the first marker 110 of the registration component in the view coordinate system and the second set of coordinate points in the world coordinate system.

[0025] In this embodiment, the first coordinate point set of the first marker 110 of the registration component in the view coordinate system and the second coordinate point set in the world coordinate system are obtained respectively. The first coordinate point set in the view coordinate system is obtained through the three-dimensional tomographic image after three-dimensional reconstruction in the above embodiment. Specifically, the three-dimensional model of the bone surface is extracted by threshold segmentation, and a view coordinate system is established. This coordinate system usually takes the upper left corner of the image or the scan center as the origin. On the surface of the reconstructed three-dimensional model, easily identifiable anatomical verification points are manually or semi-automatically selected. The selected verification points must simulate real clinical registration. The verification points are divided into two point sets, namely the registration point set. (i.e., the first set of coordinate points) and the set of marked points (i.e., the third set of coordinate points).

[0026] The second set of coordinate points in the world coordinate system can be obtained in two ways. One method is to fix the prepared tibial model on the worktable of a high-precision coordinate measuring machine (CMM), ensuring the model is stable and its orientation facilitates probe contact with all pre-embedded markers. The CMM's measurement accuracy is better than 0.05 mm (typical range 0.005-0.05 mm), and its machine coordinate system is the defined world coordinate system. The registration point set is acquired using the CMM's pointed tip, and the marker point set is acquired using its round tip, thus obtaining the registration point set in the world coordinate system. (i.e., the second set of coordinate points) and the set of marked points (Fourth coordinate point set) The contact measurement of the coordinate measuring machine is unaffected by environmental factors such as ambient light, temperature, humidity, and airflow. The stability and repeatability of the verification results are far superior to the optical positioning instrument-based scheme (the measurement accuracy of the optical positioning instrument is 0.2mm, while the measurement accuracy of the coordinate measuring machine can reach 0.005-0.05mm). Alternatively, the coordinates of the registration point set can be obtained using an optical positioning instrument (simulating actual clinical operation), and the coordinates of the marker point set can be obtained using a coordinate measuring machine (CMM) (simulating actual clinical operation). The advantage of this method is that the registration process completely simulates the clinical procedure (in clinical practice, the navigation system uses an optical positioning instrument for registration), while the verification benchmark uses high-precision measurement equipment, thus balancing clinical authenticity and verification reliability.

[0027] The set of registration points in the view coordinate system (i.e., the first set of coordinate points) and the registration point set in the world coordinate system The second set of coordinate points (i.e., the second set of coordinate points) is input into the registration algorithm to be verified, and the algorithm calculates the reference relationship matrix from image space to physical space. (Typically a 4×4 rigid body transformation matrix, including rotation and translation components).

[0028] Step S203: Calculate the mapped coordinate point set of the fourth coordinate point set in the view coordinate system based on the fourth coordinate point set of the second marker 120 of the marker component in the world coordinate system and the reference relationship matrix between the coordinate systems.

[0029] In this embodiment, the reference relationship matrix from the view coordinate system to the world coordinate system is obtained. Subsequently, this application utilizes this matrix to analyze the fourth coordinate point set. Coordinate mapping is performed due to the relation matrix. This describes the transformation between the view coordinate system and the world coordinate system, but what we need is the inverse transformation between the world coordinate system and the view coordinate system. This requires calculation. inverse matrix Thus, the mapped coordinate point set of the fourth coordinate point set in the view coordinate system is obtained. This is used to calculate the error between the view coordinate system and the view coordinate system.

[0030] Step S204: Calculate the distance error, angle error, and plane error between the third set of coordinate points of the second marker 120 of the marking component in the view coordinate system and the mapped coordinate point set, wherein the angle error is the difference in the angle between the direction vectors of at least two points in the third set of coordinate points and the mapped coordinate point set before and after registration; the plane error is the difference in the plane normal vectors between three non-collinear points in the third set of coordinate points and the mapped coordinate point set.

[0031] In this embodiment, the third set of coordinate points in the view coordinate system is obtained. With the set of mapped coordinate points Next, the errors between the coordinate point sets are evaluated, specifically including distance error, angle error, and planar error. Distance error refers to calculating the true coordinates (third coordinate point set) of each marker point in the view coordinate system. ) and the registered mapped coordinates (mapped coordinate point set) The Euclidean distance deviation between them is calculated using the following formula;

[0032] in, Indicates the first The view coordinates of the marked points Indicates the first Mapped coordinates of each marker point Quantified the first The straight-line distance deviation between the theoretical position of a marker point in the image and the position predicted by this registration directly reflects the positional accuracy of the registration at that local point, and the unit is millimeters (mm).

[0033] Based on all The individual distance errors at each point are used to calculate global statistical indicators to comprehensively evaluate the overall registration accuracy. Specifically, the root mean square error (RMSE) and repeatability error are used. RMSE is an indicator that measures the overall level of error and is more sensitive to larger errors, effectively reflecting the magnitude of the error. RMSE provides a scalar measure of the overall distance error of the entire set of marker points. A lower RMSE value indicates higher registration accuracy for the entire area. The RMSE value is calculated using the following formula:

[0034] in, It is the Euclidean distance deviation. Indicates that there is One marker point, Indicates the first 1 marker point.

[0035] Repeatability error is the error caused by repeated measurements of the same point. Next, calculate Standard deviation of Standard deviation measures the degree of dispersion of individual distance errors around the mean. Smaller values ​​indicate a more concentrated error distribution and better consistency in registration accuracy. A larger value indicates unstable registration accuracy and the possible presence of systematic bias or standard deviation. Calculated using the following formula:

[0036] in, express The average of the times, Indicates the number of times a sample was collected from the same point. Standard deviation. It should be less than the clinically acceptable registration error to be evaluated.

[0037] In addition, error calculation also includes angular error and planar error. Angular error is the difference in angle between the direction vectors of at least two points in the third coordinate point set and the mapped coordinate point set before and after registration. Angular error quantifies the inaccuracy of the rigid body rotation component during registration, reflecting the deviation of the navigation system in determining the orientation of instruments or anatomical structures. This error is determined only by the relative direction of the two points and is not affected by overall translational error. For example, in long bone surgery, it directly corresponds to the deviation of the force line direction of the intramedullary nail or osteotomy guide plate, compensating for the deficiency that simple distance error cannot assess rotational deviation. For surgeries with high angle requirements, such as spinal correction and joint replacement, angular error is a key performance indicator for determining the usability of the navigation system. Planar error is the difference in the plane normal vector between three non-collinear points in the third coordinate point set and the mapped coordinate point set. This vector difference can be used to analyze the specific tilt component of the plane in a specific anatomical direction (such as anterior-posterior, lateral-lateral). By evaluating the error distribution of the plane in different regions, the uniformity of registration accuracy in three-dimensional space can be determined, and the existence of local distortions can be identified.

[0038] Step S205: Verify the accuracy of registration of the three-dimensional tomographic image acquired by the first image acquisition device based on the distance error, angle error, and planar error.

[0039] In summary, this application obtains registration point sets and marker point sets in both the view coordinate system and the world coordinate system by non-uniformly embedding verification points in the bone model. It then calculates the reference relationship matrix between the two coordinate systems to obtain the marker point set and mapped coordinate point set in the view coordinate system. By calculating the distance error, angle error, and plane error between these two sets, the registration accuracy is verified. This solves the problem of accuracy averaging caused by uniform distribution of verification points, ensuring that the assessment conclusions can accurately predict risks in core clinical areas. By introducing plane error (assessing changes in the plane normal) and angle error (assessing changes in the axial direction), it overcomes the limitation of existing technologies that can only assess distance errors. This fills the technical gap in verification methods for surgeries with stringent requirements for angular accuracy, such as joint replacement and spinal correction. It provides an objective, reliable, and comprehensive verification method for the registration accuracy of surgical navigation systems, improving the reliability and versatility of registration accuracy verification in surgical navigation technology.

[0040] In one implementation, such as Figure 3 As shown, in step S204, the angle error is the difference in angle between the direction vectors of at least two points in the third coordinate point set and the mapped coordinate point set before and after registration, including: In this embodiment, the set of marker points in the view coordinate system (i.e., the third coordinate point set) is used. The straight line formed by two points and the set of mapped coordinate points in the equation () The angle between the straight lines formed by the two corresponding points is calculated to obtain the angle difference between the direction vectors before and after registration. The specific calculation method is as follows.

[0041] Step S301, as follows Figure 4 As shown, the angle difference between the direction vectors before and after registration can be calculated by selecting the first and second marker points from the third coordinate point set.

[0042] In this embodiment, in the third coordinate point set Select the first and second marker points, and calculate the vector formed by the first and second marker points and the mapped coordinate set. The angle between the vectors formed by two corresponding points in the model is denoted as the two-point method, which directly quantizes the angular deviation during the registration process. The specific calculation formula is as follows:

[0043] in, Here is the formula for calculating the angle between vectors. This represents the vector representing the first and second marker points in the view coordinate system. This represents the vector formed by the corresponding positions of the first and second marked points in the mapped coordinate point set. This represents the formula for the dot product of vectors. Representing vectors The modulus formula, Representing vectors The formula for the modulus.

[0044] In this embodiment, the two-point method is the basis for angular error assessment. It is usually used to verify the orientation accuracy of specific and critical anatomical axes (such as the femoral mechanical axis and screw planning trajectory). In multi-point pair assessments, it serves as a standard function for calculating individual angular error values.

[0045] Step S302, the angle difference between the direction vectors before and after registration can also be calculated by selecting at least two marker points with axial distribution in the third coordinate point set.

[0046] In this embodiment, by selecting at least two marker points distributed along a specific anatomical axis (such as the sagittal axis of the spine or the mechanical axis of long bones), an axis direction vector is constructed or fitted, and the angle between the directions of the axes before and after registration is compared to evaluate the registration accuracy; this is defined as the axis method. In the third coordinate point set... In the target axis, a set of p points (p ≥ 2) are distributed along the target axis. For example, to evaluate the tibial mechanical axis, multiple points near the center of the tibial plateau and the center of the ankle joint are selected, and their direction vectors are calculated. The coordinates are then mapped to the set of points. Find the points corresponding to the selected p points to obtain the direction vector. Use the formula for calculating the included angle using the two-point method in step S301 to obtain the angle error. This scheme is suitable for evaluating the rotational registration accuracy of linear or columnar structures such as the spine and long bones, assisting the registration algorithm and providing optimized directions.

[0047] In one embodiment, the angle error is the difference in the angle between the direction vectors of the third coordinate point set and at least two points in the mapped coordinate point set before and after registration, and further includes: the difference in the angle between the direction vectors before and after registration can also be calculated by selecting at least two pairs of marker points in the third coordinate point set.

[0048] Based on the two-point method in the above embodiments, a multi-point averaging method can be extended. This method selects multiple different pairs of marked points, calculates the rotation angle error of each pair, and statistically analyzes these errors to obtain a more stable overall rotation angle assessment. In the third coordinate point set... In this study, multiple different and representative point pair combinations are selected, such as point pairs (1,2), (2,3), (1,3), etc. For each point pair, the angle error of each pair is calculated independently using the two-point method. A statistical analysis is performed on the calculated set of angle errors, and the average or maximum value is used as the final angle error evaluation index, calculated using the following formula:

[0049]

[0050] in, Indicates the first Angular error of the pair of points This indicates the number of pairs of points selected. This represents the average angle error of multiple point pairs. This represents the maximum angular error among multiple point pairs. The average or maximum value is taken as the final evaluation metric.

[0051] Average angular error The maximum angular error is the core evaluation indicator for overall rotational accuracy. For the critical value of rotational accuracy, when Significantly greater than For example, if the difference is greater than the threshold of 2° (the specific threshold is set according to the actual situation), it indicates that there is a large rotation error in a local area or a specific direction, which may indicate local registration failure or abnormal marker points, and adjustments are required.

[0052] This embodiment employs multiple methods for solving angular errors, applicable to different surgical scenarios. It expands the evaluation dimensions of registration error from "point" to "vector / plane / axis", making the verification system more comprehensive and scientific.

[0053] In one embodiment, the plane error in step S204 is the difference in plane normal vectors between the third coordinate point set and three non-collinear points in the mapped coordinate point set, and may include: Based on the reference plane formed by the three non-collinear points in the third coordinate point set, the normal angle and tilt direction of the reference plane before and after registration are calculated and visualized.

[0054] In this embodiment, a reference plane is formed by selecting three non-collinear verification points in the view coordinate system. The spatial attitude changes of this plane before and after registration are compared, and its normal vector angle and tilt direction are calculated. For example, a third set of coordinate points is selected. Given three points P1, P2, and P3, calculate the center point among them:

[0055] in, , , Let P1, P2, and P3 represent the coordinates of the three points respectively. Represents the coordinates of the center point.

[0056] The normal vector of the plane formed by points P1, P2, and P3 is calculated using the following formula:

[0057]

[0058] , in, Let P1 be the vector pointing from point P1 to point P2. Let P1 be the vector pointing from point P1 to point P3. Indicates the direction perpendicular to the vector and The plane in which it lies, i.e., the normal vector of the plane in question. Representing vectors and The magnitude of the cross product vector, Let represent the unit normal vector of the plane in question. (Based on the unit normal vector of the plane) The local orthogonal basis vectors in the plane (based on the reference vector) are calculated using the following formula. (To prevent parallelism with the normal)

[0059]

[0060] in, This represents a reference vector used to assist in constructing basis vectors in the plane; the default value is... ,like , then change to , Indicates a perpendicular to and Vectors in the plane, Representing vectors and The magnitude of the cross product vector is used for normalization. Let represent the first unit basis vector in the plane. Indicates a perpendicular to and Vectors in the plane, Representing vectors and The magnitude of the cross product vector, the vector Let the second unit basis vector in the plane be denoted by the vector . Orthogonal.

[0061] The projection radius of three points in a plane can be calculated using the following formula based on orthogonal basis vectors:

[0062]

[0063] in, Indicates the first Marker points ( ), Indicates the first The coordinates of the marker points Indicates the coordinates of the center point. Indicates the marked point among the three points. To the center point The projected distance in the direction is taken as The point with the largest projected distance in the direction is denoted as the projection radius. . Indicates the marked point among the three points. To the center point The projected distance in the direction is taken as The point with the largest projected distance in the direction is denoted as the projection radius. .

[0064] Calculate the half-side length of the rectangle based on the projection radius (increase the side margin by 20%, and set a lower limit; the 20% is adjustable):

[0065]

[0066] in, Indicates the original projection radius Expand outwards by 20% (increase the margin by 20%). Indicates the original projection radius Expand outwards by 20% from the base value; 0.1 is the minimum lower limit for the half-side length of the rectangle to prevent numerical instability or missed detections due to excessively small dimensions. express The length of half the rectangle in the direction, express The length of half a rectangle in the direction.

[0067] Calculate the half-side length of a rectangle and calculate the four vertices of the rectangle based on the projection radius:

[0068]

[0069]

[0070]

[0071] in,( , , , The ) represent the four vertices of the rectangle in the view coordinate system. Indicates the coordinates of the center point. express The length of half the rectangle in the direction, express The length of the rectangle's half-side in the direction, arranged in the order of the four vertices above, forms a quadrilateral Q1.

[0072] For the set of mapped coordinate points The corresponding positions of points P1, P2, and P3 in the middle Repeat the above steps at three points to obtain another quadrilateral. The angle between the normals and the tilt direction angle are calculated using two plane normal vectors.

[0073] Visualizing quadrilaterals, such as Figure 5As shown, the relative positions and angles of the two planes can be seen intuitively. Error information is obtained based on the visualization, and an optimization strategy for the error information is generated. The optimization strategy is used to adjust the execution mode of the subsequent registration method. For example, if the image shows that the image always shifts to the left after each registration verification, then it is possible to consider adding registration points in other areas, or to perform data processing on the registration algorithm. The specific operation method needs to be optimized according to the specific offset situation, and the registration algorithm corresponding to a certain situation needs to be optimized.

[0074] This embodiment provides optimization directions for the registration algorithm. If multiple verifications show a systematic offset in the same direction, the weight of the registration points can be increased in the corresponding area, or a targeted registration strategy can be designed to improve the registration accuracy.

[0075] Figure 6 This is a schematic diagram of a registration accuracy verification device according to an embodiment of this application. Figure 6 As shown, the device includes the following units: The image acquisition unit 601 is used to acquire three-dimensional tomographic images of the surgical area of ​​the bone model acquired by the registration and marking components of the first image acquisition device.

[0076] The coordinate transformation unit 602 is used to calculate the reference relationship matrix between the coordinate systems based on the first set of coordinate points of the first marker of the registration component in the view coordinate system and the second set of coordinate points in the world coordinate system.

[0077] The coordinate acquisition unit 603 is used to calculate the mapped coordinate point set of the fourth coordinate point set in the view coordinate system based on the fourth coordinate point set of the second marker of the marker component in the world coordinate system and the reference relationship matrix between the coordinate systems.

[0078] The error calculation unit 604 is used to calculate the distance error, angle error, and plane error between the third set of coordinate points of the second marker of the marking component in the view coordinate system and the mapped coordinate point set, wherein the angle error is the difference in the angle between the direction vectors of at least two points in the third set of coordinate points and the mapped coordinate point set before and after registration; the plane error is the difference in the plane normal vector between three non-collinear points in the third set of coordinate points and the mapped coordinate point set.

[0079] The verification unit 605 is used to verify the accuracy of the registration of the three-dimensional tomographic images acquired by the first image acquisition device based on the distance error, angle error, and plane error.

[0080] In one implementation, such as Figure 7 As shown, the error calculation unit 604 may further include the following modules: The first calculation module 701 is used to calculate the angle difference between the direction vectors before and after registration by selecting the first and second marker points from the third coordinate point set.

[0081] The second calculation module 702 is used to calculate the angle difference between the direction vector before and after registration by selecting at least two marker points with axis distribution in the third coordinate point set.

[0082] In summary, this application's embodiments obtain registration point sets and marker point sets in both the view coordinate system and the world coordinate system by non-uniformly embedding verification points in the bone model. The reference relationship matrix between the two coordinate systems is then calculated to obtain the marker point set and mapped coordinate point set in the view coordinate system. By calculating the distance error, angle error, and plane error between these two sets, the registration accuracy is verified. This solves the problem of accuracy averaging caused by uniform distribution of verification points, ensuring that the evaluation conclusions can accurately and sensitively predict risks in core clinical areas. By introducing plane error (assessing changes in the plane normal) and angle error (assessing changes in the axial direction), the limitations of existing technologies that can only assess distance error are overcome. This fills the technical gap in verification methods for surgeries with stringent requirements for angular accuracy, such as joint replacement and spinal correction. It provides an objective, reliable, and comprehensive verification method for the registration accuracy of surgical navigation systems, improving the reliability and versatility of registration accuracy verification in surgical navigation technology.

[0083] It should be noted that those skilled in the art will understand that the different implementation methods, their descriptions and explanations, and the technical effects achieved as described in the method embodiments of this application are also applicable to the device embodiments of this application, and will not be repeated here.

[0084] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0085] The foregoing description of exemplary embodiments of this application should be understood as not limiting, but illustrative, and the scope of protection of this application is not limited thereto. It should be understood that those skilled in the art can make modifications and variations to the embodiments of this application without departing from the spirit and scope of this application, and such modifications and variations should be within the scope of protection of this application.

Claims

1. A registration accuracy verification method, characterized in that, The method includes: Acquire three-dimensional tomographic images of the surgical area of ​​the bone model using the registration and marking components of the first image acquisition device; The reference relationship matrix between the coordinate systems is calculated based on the first set of coordinate points of the first marker in the view coordinate system and the second set of coordinate points in the world coordinate system of the registration component. The mapped coordinate point set of the fourth coordinate point set in the view coordinate system is calculated based on the fourth coordinate point set of the second marker of the marker component in the world coordinate system and the reference relationship matrix between the coordinate systems; Calculate the distance error, angle error, and planar error between the third set of coordinate points of the second marker of the marking component in the view coordinate system and the mapped coordinate point set, wherein the angle error is the difference in the angle between the direction vectors of at least two points in the third set of coordinate points and the mapped coordinate point set before and after registration; the planar error is the difference in the plane normal vectors between three non-collinear points in the third set of coordinate points and the mapped coordinate point set. The accuracy of the registration of the three-dimensional tomographic images acquired by the first image acquisition device is verified based on the distance error, angle error, and planar error.

2. The registration accuracy verification method according to claim 1, characterized in that, The angle error is the difference in angle between the direction vectors of at least two points in the third coordinate point set and the mapped coordinate point set before and after registration, including: The angle difference between the direction vectors before and after registration can be calculated by selecting the first and second marker points in the third coordinate point set. Alternatively, the angle difference between the direction vectors before and after registration can be calculated by selecting at least two marker points with axial distribution from the third coordinate point set.

3. The registration accuracy verification method according to claim 1, characterized in that, The angle error is the difference in angle between the direction vectors of at least two points in the third coordinate point set and the mapped coordinate point set before and after registration, including: The angle difference between the direction vectors before and after registration can also be calculated by selecting at least two pairs of marker points from the third coordinate point set.

4. The registration accuracy verification method according to claim 3, characterized in that, The angle difference between the direction vectors before and after registration can also be calculated by selecting at least two pairs of marker points from the third coordinate point set, including: Calculate the angle difference between at least two pairs of marked points, and take the average or maximum value as the final evaluation index.

5. The registration accuracy verification method according to claim 1, characterized in that, The plane error is the difference in plane normal vectors between the third set of coordinate points and three non-collinear points in the mapped set of coordinate points, including: Based on the reference plane formed by the three non-collinear points in the third coordinate point set, the normal angle and tilt direction of the reference plane before and after registration are calculated and visualized.

6. The registration accuracy verification method according to claim 5, characterized in that, The calculation of the normal angle and tilt direction of the reference plane before and after registration, and its visualization, includes: Based on the error information obtained from the visualization, an optimization strategy for the error information is generated. The optimization strategy is used to adjust the execution mode of the subsequent registration method.

7. The registration accuracy verification method according to claim 1, characterized in that, The second markers of the marking component are arranged in a locally dense, non-uniform distribution, wherein the marking density in the surgical area and the edge away from the registration component is higher than that in other areas.

8. A registration accuracy verification device, characterized in that, The device includes: The image acquisition unit is used to acquire three-dimensional tomographic images of the surgical area of ​​the bone model acquired by the registration and marking components of the first image acquisition device. The coordinate transformation unit is used to calculate the reference relationship matrix between the coordinate systems based on the first set of coordinate points of the first marker of the registration component in the view coordinate system and the second set of coordinate points in the world coordinate system. The coordinate acquisition unit is used to calculate the mapped coordinate point set of the fourth coordinate point set in the view coordinate system based on the fourth coordinate point set of the second marker of the marker component in the world coordinate system and the reference relationship matrix between the coordinate systems; The error calculation unit is used to calculate the distance error, angle error, and planar error between the third set of coordinate points of the second marker of the marking component in the view coordinate system and the mapped coordinate point set, wherein the angle error is the difference in the angle between the direction vectors of at least two points in the third set of coordinate points and the mapped coordinate point set before and after registration; the planar error is the difference in the plane normal vectors between three non-collinear points in the third set of coordinate points and the mapped coordinate point set. The verification unit is used to verify the accuracy of the registration of the three-dimensional tomographic images acquired by the first image acquisition device based on the distance error, angle error, and plane error.

9. An electronic device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it causes the electronic device to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images