A device for determining posterior inclination angle of tibial plateau

Abstract

The present application provides a device for determining the posterior inclination angle of a tibial plateau, the determination device comprising: an acquisition module for acquiring a first plane where the lowest point of the tibial plateau in a three-dimensional tibial image is located; the first plane is parallel to the cross section of the three-dimensional tibial image; a first determination module for determining a target area containing the tibial plateau and an intersection line corresponding to the first plane; a second determination module for rotating the first plane along a preset direction M times by a preset angle with the intersection line as the axis to obtain M rotation planes; a calculation module for calculating, for each rotation plane, the sum of the Mahalanobis distances of all points in the target area to the rotation plane; and a third determination module for determining the angle between the rotation plane with the smallest sum of the Mahalanobis distances and the first plane as the posterior inclination angle of the tibial plateau. The present application solves the technical problem of inaccurate determination of the posterior inclination angle of the tibial plateau by identifying two-dimensional images in the prior art, and achieves the technical effect of improving the accuracy of the posterior inclination angle of the tibial plateau.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a device for determining the posterior inclination angle of a tibial plateau. Background Art

[0002] The posterior tibial slope is the tilt angle of the tibial plateau, an important structure of the knee joint morphology, and is often used clinically to evaluate the success of total knee surgery. Furthermore, obtaining an accurate posterior tibial slope during preoperative surgical planning is a key step in the success of surgery.

[0003] In the prior art, the posterior inclination angle of the tibial plateau is mostly measured through 2D X-rays, which results in inaccurate measurement of the posterior inclination angle of the tibial plateau. Summary of the invention

[0004] In view of this, the present application provides a device for determining the posterior inclination angle of the tibial plateau, which solves the technical problem of inaccurate identification of the posterior inclination angle of the tibial plateau by two-dimensional image recognition in the prior art by determining the posterior inclination angle of the tibial plateau, and achieves the technical effect of improving the accuracy of the posterior inclination angle of the tibial plateau.

[0005] An embodiment of the present application provides a device for determining the posterior inclination angle of a tibial plateau, and the device for determining the posterior inclination angle of a tibial plateau includes: an acquisition module, used to acquire a first plane where the lowest point of the tibial plateau in a three-dimensional tibial image is located; the first plane is parallel to the cross-section of the three-dimensional tibial image; a first determination module, used to determine a target area containing the tibial plateau, and an intersection line corresponding to the first plane; a second determination module, used to rotate the first plane M times by a preset angle along a preset direction with the intersection line as the axis to obtain M rotation planes; a calculation module, used to calculate, for each rotation plane, the sum of the Mahalanobis distances from all points in the target area to the rotation plane; and a third determination module, used to determine the angle between the rotation plane with the smallest sum of the Mahalanobis distances and the first plane as the posterior inclination angle of the tibial plateau.

[0006] Optionally, the acquisition module obtains the first plane where the lowest point of the tibial plateau in the three-dimensional tibial image is located through the following steps: inputting the three-dimensional tibial image into a trained neural network model to obtain the lowest point of the tibial plateau predicted by the neural network model; the lowest point is the point with the lowest height in the tibial plateau.

[0007] Optionally, the first determination module includes: a prediction submodule, used to input the three-dimensional tibial image into a trained neural network model to obtain the lateral points predicted by the neural network model; a segmentation submodule, used to segment the three-dimensional tibial image with a second plane containing the lateral points; wherein the second plane is parallel to the cross-section of the three-dimensional tibial image and is at a preset distance from the first plane; and a selection submodule, used to select the segmented three-dimensional tibial image containing the tibial plateau as the target area.

[0008] Optionally, the first determination module also includes: a vertical plane determination submodule, used to determine a vertical plane perpendicular to the first plane, wherein the vertical plane includes boundary points in the first plane; and an intersection line determination submodule, used to determine the intersection line between the vertical plane and the first plane as the intersection line corresponding to the first plane.

[0009] Optionally, the vertical plane determination submodule includes: a first determination unit, used to determine all intersection points of the first plane and the three-dimensional tibial image; a second determination unit, used to determine a straight line perpendicular to a line connecting the lowest points of the tibial plateau, the straight line including the midpoint of the line connecting the lowest points of the tibial plateau; a third determination unit, used to determine, among the points where the straight line coincides with all the intersection points, a point located on the front side of the three-dimensional tibial image as a boundary point; and a fourth determination unit, used to determine a vertical plane including the boundary point and perpendicular to the first plane.

[0010] Optionally, the neural network model is trained by the following steps: all points in the three-dimensional tibial image are used as training samples; the lowest point and side points of the tibial plateau manually marked in the three-dimensional tibial image are used as labels to train the neural network model; the side points are points located on the side of the three-dimensional tibial image.

[0011] Optionally, the loss function of the neural network model is calculated as:

[0012]

[0013] The calculation formula of the activation function of the neural network model is:

[0014]

[0015] In the above formula, L refers to the loss function of the neural network model, N refers to the number of training samples, and α i refers to the i-th training sample, y i Refers to the α output by the neural network model i The corresponding predicted probability, f(α i ) refers to the i-th activation function value.

[0016] Optionally, the neural network model includes 5 convolution layers and 5 pooling layers, each convolution layer includes two 3×3×3 convolution kernels; each pooling layer includes a 2×2×2 pooling kernel.

[0017] Optionally, the calculation module calculates the Mahalanobis distances from all points in the target area to the rotation plane by the following formula:

[0018]

[0019] In the above formula, a j refers to the coordinate vector of the jth point in the target area, b j Refers to the coordinate vector of the projection point of the jth point in the target area on the rotation plane, D m (a j ) refers to the Mahalanobis distance from the jth point in the target area to the rotation plane, Σ -1 Refers to the inverse matrix of the covariance matrix of the target region.

[0020] Optionally, the preset angle is 0.5 degrees.

[0021] An embodiment of the present application provides a device for determining the posterior inclination angle of a tibial plateau, and the device for determining the posterior inclination angle of a tibial plateau includes: an acquisition module, used to acquire a first plane where the lowest point of the tibial plateau in a three-dimensional tibial image is located; the first plane is parallel to the cross-section of the three-dimensional tibial image; a first determination module, used to determine a target area containing the tibial plateau, and an intersection line corresponding to the first plane; a second determination module, used to rotate the first plane M times in a preset direction by a preset angle with the intersection line as the axis to obtain M rotation planes; a calculation module, used to calculate, for each rotation plane, the sum of the Mahalanobis distances from all points in the target area to the rotation plane; and a third determination module, used to determine the angle between the rotation plane with the smallest sum of the Mahalanobis distances and the first plane as the posterior inclination angle of the tibial plateau. By determining the rotation plane with the smallest sum of Mahalanobis distances among various rotation planes of the three-dimensional tibial image, and determining the angle between the rotation plane with the smallest sum of Mahalanobis distances and the first plane as the posterior inclination angle of the tibial plateau, the technical problem of inaccurate identification of the posterior inclination angle of the tibial plateau by the prior art of two-dimensional image recognition is solved, and the technical effect of improving the accuracy of the posterior inclination angle of the tibial plateau is achieved.

[0022] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 A functional module diagram of a device for determining a tibial plateau posterior inclination angle provided in an embodiment of the present application is shown.

[0025] Figure 2 A schematic diagram of a three-dimensional bone image provided in an embodiment of the present application is shown.

[0026] Figure 3 A schematic diagram showing the lowest point of the tibial plateau provided by an embodiment of the present application is shown.

[0027] Figure 4 A schematic diagram showing the side points of the tibial plateau provided by an embodiment of the present application is shown.

[0028] Figure 5 A top view of a first plane provided by an embodiment of the present application is shown. DETAILED DESCRIPTION

[0029] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art, under the guidance of the content of the present application, can add one or more other operations to the flowchart, or remove one or more operations from the flowchart.

[0030] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0031] In the prior art, the tibial plateau inclination angle is measured by using a 2D X-ray film, which results in an inaccurate measurement of the tibial plateau inclination angle.

[0032] Based on this, the embodiment of the present application provides a device for determining the posterior inclination angle of the tibial plateau, which solves the technical problem of inaccurate determination of the posterior inclination angle of the tibial plateau by identifying two-dimensional images in the prior art by determining the rotation plane with the smallest sum of Mahalanobis distances in each rotation plane of the three-dimensional tibial image, and determines the angle between the rotation plane with the smallest sum of Mahalanobis distances and the first plane as the posterior inclination angle of the tibial plateau, thereby achieving the technical effect of improving the accuracy of the posterior inclination angle of the tibial plateau. The details are as follows:

[0033] See also Figure 1 , Figure 1 This is a schematic diagram of a module of a device for determining the posterior inclination angle of a tibial plateau provided in an embodiment of the present application. Figure 1 As shown, the device for determining the posterior inclination angle of the tibial plateau includes: an acquisition module 101 , a first determination module 102 , a second determination module 103 , a calculation module 104 and a third determination module 105 .

[0034] The acquisition module 101 is used to acquire a first plane where the lowest point of the tibial plateau in the three-dimensional tibial image is located; the first plane is parallel to the cross section of the three-dimensional tibial image.

[0035] The three-dimensional tibia image is obtained in the three-dimensional bone image, wherein the three-dimensional bone image includes a three-dimensional femur image and a three-dimensional tibia image. Figure 2 , Figure 2 Schematic diagram of a three-dimensional bone image provided in an embodiment of the present application. Figure 2 As shown, the three-dimensional femur image 1 is located in the upper half of the three-dimensional bone image, and the three-dimensional tibia image 2 is located in the lower half of the three-dimensional bone image.

[0036] Obtaining a three-dimensional tibia image from a three-dimensional bone image includes: inputting the three-dimensional bone image into a pre-trained neural network model to obtain a three-dimensional tibia image predicted by the neural network model. That is, the three-dimensional bone image is used as a training sample, and the three-dimensional tibia image manually marked in the three-dimensional bone image is used as a label to train the neural network model.

[0037] The three-dimensional tibia image is of the tibia of a human body standing on the horizontal ground, that is, the cross section of the three-dimensional tibia image is parallel to the horizontal ground.

[0038] The acquisition module acquires the first plane where the lowest point of the tibial plateau in the three-dimensional tibial image is located through the following steps: inputting the three-dimensional tibial image into a trained neural network model to obtain the lowest point of the tibial plateau predicted by the neural network model; the lowest point is the point with the lowest height in the tibial plateau.

[0039] See also Figure 3 , Figure 3 Schematic diagram of the lowest point of the tibial plateau provided in the embodiment of the present application. Figure 3 As shown, the coordinate system of the three-dimensional tibia image can be set to take the cross section of the three-dimensional tibia image as the XOY plane, and take the upward direction of the cross section of the three-dimensional tibia image as the Z axis of the coordinate system. The number of the lowest points 3 of the tibia plane is two, and in the embodiment of the present application, the first plane is a plane containing the lowest points of the two tibia planes and the cross-sectional plane of the three-dimensional tibia image.

[0040] The first determination module 102 is used to determine a target area including the tibial plateau and an intersection line corresponding to the first plane.

[0041] The first determination module 102 includes: a prediction submodule, which is used to input the three-dimensional tibial image into the trained neural network model to obtain the side points predicted by the neural network model; a segmentation submodule, which is used to segment the three-dimensional tibial image with a second plane containing the side points; wherein the second plane is parallel to the cross-section of the three-dimensional tibial image and is at a preset distance from the first plane; and a selection submodule, which is used to select the segmented three-dimensional tibial image containing the tibial plateau as the target area.

[0042] The preset distance is usually set to 5 cm. That is, the upper part of the second plane in the 3D tibia image is used as the target area. Figure 4 , Figure 4 Schematic diagram of the side points of the tibial plateau provided in the embodiment of the present application. Figure 4 As shown, the lateral point 4 is a point located on the lateral side of the three-dimensional tibia image. Figure 3 and Figure 4 The 3D tibia image in is the target area.

[0043] The neural network model is trained through the following steps: all points in the three-dimensional tibial image are used as training samples; the lowest point and side points of the tibial plateau manually marked in the three-dimensional tibial image are used as labels to train the neural network model.

[0044] The calculation formula of the loss function of the neural network model is:

[0045]

[0046] The calculation formula of the activation function of the neural network model is:

[0047]

[0048] In formulas (1) and (2), L refers to the loss function of the neural network model, N refers to the number of training samples, and α irefers to the i-th training sample, y i Refers to the α output by the neural network model i The corresponding predicted probability, f(α i ) refers to the i-th activation function value.

[0049] The neural network model consists of 5 convolutional layers and 5 pooling layers. Each convolutional layer contains two 3×3×3 convolution kernels; each pooling layer contains a 2×2×2 pooling kernel.

[0050] The first determination module 102 also includes: a vertical plane determination submodule, used to determine a vertical plane perpendicular to the first plane, wherein the vertical plane includes boundary points in the first plane; and an intersection line determination submodule, used to determine the intersection line between the vertical plane and the first plane as the intersection line corresponding to the first plane.

[0051] The vertical plane determination submodule includes: a first determination unit, which is used to determine all intersection points of the first plane and the three-dimensional tibial image; a second determination unit, which is used to determine a straight line perpendicular to a line connecting the lowest points of the tibial plateau, wherein the straight line includes the midpoint of the line connecting the lowest points of the tibial plateau; a third determination unit, which is used to determine a point located on the front side of the three-dimensional tibial image as a boundary point among the points where the straight line coincides with all the intersection points; and a fourth determination unit, which is used to determine a vertical plane including the boundary point and perpendicular to the first plane.

[0052] See also Figure 5 , Figure 5 This is a top view of the first plane provided in the embodiment of the present application. Figure 5 As shown, straight line L1 refers to the line connecting the lowest points of the tibial plateau (i.e., both ends of straight line L1 are the lowest points of the tibial plateau), straight line L2 is the midpoint of the line connecting the lowest points of the tibial plateau and is perpendicular to the line connecting the lowest points of the tibial plateau, straight line L1 and straight line L2 are both in the first plane, and among the points where straight line L2 coincides with all the intersection points, the point located on the front side of the three-dimensional tibial image is determined as boundary point 5. Boundary point 5 is at the same position height as the lowest point of the tibial plateau, that is, the Z-axis coordinate value of boundary point 5 is the same as the Z-axis coordinate value of the lowest point of the tibial plateau. Figure 5 As shown, the point with the largest Y-axis coordinate value is selected from the points where the straight line overlaps with all the intersection points, and is determined as the point located at the front side of the three-dimensional tibia image. In other words, the point located at the front side of the three-dimensional tibia image is determined based on the establishment method of the coordinate system and the coordinates of the points where the straight line overlaps with all the intersection points. The plane containing the boundary point 5 and perpendicular to the first plane is determined as the vertical plane, and the intersection line of the vertical plane and the first plane is determined as the intersection line corresponding to the first plane.

[0053] The second determining module 103 is configured to rotate the first plane M times by a preset angle along a preset direction with the intersection line as the axis to obtain M rotation planes.

[0054] The preset direction is clockwise, M times is generally set to 60 times, and the preset angle is generally set to 0.5 degrees. That is, the first plane is rotated 60 times clockwise with the intersection line as the axis, each time rotating 0.5 degrees, to obtain a rotation plane corresponding to each rotation.

[0055] The calculation module 104 is used to calculate, for each rotation plane, the sum of the Mahalanobis distances from all points in the target area to the rotation plane.

[0056] The calculation module calculates the Mahalanobis distance from all points in the target area to the rotation plane using the following formula:

[0057]

[0058] In formula (3), a j refers to the coordinate vector of the jth point in the target area, b j Refers to the coordinate vector of the projection point of the jth point in the target area on the rotation plane, D m (a j ) refers to the Mahalanobis distance from the jth point in the target area to the rotation plane, Σ -1 Refers to the inverse matrix of the covariance matrix of the target region.

[0059] For example, The coordinate vector a of the jth point in the target area j is the horizontal coordinate x of the jth point j , the vertical coordinate y j and the position height coordinate z j The column vector composed of The coordinate vector b of the projection point of the jth point in the target area on the rotation plane j It is a column vector consisting of the horizontal coordinate x0, vertical coordinate y0 and position height coordinate z0 of the projection point.

[0060] Specifically, the method for determining the projection point of the j-th point in the target area on the rotation plane is: determine a straight line that includes the j-th point in the target area and is perpendicular to the rotation plane, and determine the intersection of the straight line and the rotation plane as the projection point of the j-th point in the target area on the rotation plane.

[0061] The covariance matrix of the target area is calculated by the following formula:

[0062]

[0063]

[0064] In formula (4), Σ refers to the covariance matrix of the target area, which refers to the covariance matrix between the horizontal coordinates, vertical coordinates and position height coordinates of all points in the target area; n refers to the number of all points in the target area; COV(X,X) refers to the covariance between the horizontal coordinates and the horizontal coordinates of all points in the target area; COV(X,Y) refers to the covariance between the horizontal coordinates and the vertical coordinates of all points in the target area; COV(X,Z) refers to the covariance between the horizontal coordinates and the vertical coordinates of all points in the target area; COV(X,Y) refers to the covariance between the horizontal coordinates and the position height coordinates of all points in the target area; C OV(Y,X) refers to the covariance of the ordinate and the abscissa of all points in the target area; COV(Y,Y) refers to the covariance of the ordinate and the ordinate of all points in the target area; COV(Y,Z) refers to the covariance of the ordinate and the position height coordinate of all points in the target area; COV(Z,X) refers to the covariance of the position height coordinate and the abscissa of all points in the target area; COV(Z,Y) refers to the covariance of the position height coordinate and the ordinate of all points in the target area; COV(Z,Z) refers to the covariance of the position height coordinate and the ordinate of all points in the target area.

[0065] In formula (5), x j Refers to the horizontal coordinate of the jth point in the target area; Refers to the average value of the horizontal coordinates of all points in the target area; j Refers to the ordinate of the jth point in the target area; It refers to the average value of the ordinate coordinates of all points in the target area. The calculation methods of COV(X,X), COV(X,Z), COV(Y,X), COV(Y,Y), COV(Y,Z), COV(Z,X), COV(Z,Y), and COV(Z,Z) refer to the calculation method of COV(X,Y) and will not be repeated here.

[0066] The third determination module 105 is used to determine the angle between the rotation plane with the smallest sum of Mahalanobis distances and the first plane as the tibial plateau tilt angle.

[0067] That is to say, for each rotation plane, the Mahalanobis distances from all points in the target area to the rotation plane are calculated, the Mahalanobis distances from all points in the target area to the rotation plane are summed, the rotation plane with the smallest sum is determined as the target plane, and the angle between the target plane and the first plane is determined as the posterior inclination angle of the tibial plateau.

[0068] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. In the several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0069] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0070] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0071] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0072] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A device for determining the posterior inclination angle of a tibial plateau, characterized in that: The device for determining the posterior inclination angle of the tibial plateau comprises: An acquisition module, used for acquiring a first plane where the lowest point of the tibial plateau in the three-dimensional tibial image is located; the first plane is parallel to the cross section of the three-dimensional tibial image; A first determination module is used to determine a target area including the tibial plateau and an intersection line corresponding to the first plane; A second determination module is used to rotate the first plane along a preset direction M times by a preset angle with the intersection line as the axis to obtain M rotation planes; A calculation module, used for calculating, for each rotation plane, the sum of Mahalanobis distances from all points in the target area to the rotation plane; A third determination module is used to determine the angle between the rotation plane with the smallest sum of the Mahalanobis distances and the first plane as the tibial plateau tilt angle; The first determination module includes: a prediction submodule, used for inputting the three-dimensional tibia image into a trained neural network model to obtain the side points predicted by the neural network model; a segmentation submodule, used for segmenting the three-dimensional tibia image with a second plane containing the side points; wherein the second plane is parallel to the cross section of the three-dimensional tibia image and is at a preset distance from the first plane; a selection submodule, used for selecting the segmented three-dimensional tibia image containing the tibial plateau as the target area; The first determination module further includes: a vertical plane determination submodule, used to determine a vertical plane perpendicular to the first plane, wherein the vertical plane includes boundary points in the first plane; an intersection line determination submodule, used to determine an intersection line between the vertical plane and the first plane as an intersection line corresponding to the first plane; The vertical plane determination submodule includes: a first determination unit, used to determine all intersection points between the first plane and the three-dimensional tibial image; a second determination unit, used to determine a straight line perpendicular to a line connecting the lowest points of the tibial plateau, wherein the straight line includes the midpoint of the line connecting the lowest points of the tibial plateau; a third determination unit, used to determine a point located on the front side of the three-dimensional tibial image as a boundary point among the points where the straight line coincides with all the intersection points; and a fourth determination unit, used to determine a vertical plane including the boundary point and perpendicular to the first plane.

2. The device for determining the posterior inclination angle of the tibial plateau according to claim 1, characterized in that: The acquisition module acquires the first plane where the lowest point of the tibial plateau in the three-dimensional tibial image is located by the following steps: The three-dimensional tibia image is input into the trained neural network model to obtain the lowest point of the tibial plateau predicted by the neural network model; the lowest point is the point with the lowest height in the tibial plateau.

3. The device for determining the posterior inclination angle of the tibial plateau according to claim 1 or 2, characterized in that: The neural network model is trained by the following steps: Taking all points in the three-dimensional tibia image as training samples; The lowest point and side points of the tibial plateau manually marked in the three-dimensional tibial image are used as labels to train the neural network model; the side points are points located on the side of the three-dimensional tibial image.

4. The device for determining the posterior inclination angle of the tibial plateau according to claim 3, characterized in that: The calculation formula of the loss function of the neural network model is: The calculation formula of the activation function of the neural network model is: In the above formula, L refers to the loss function of the neural network model, N refers to the number of training samples, and α i refers to the i-th training sample, y i Refers to the α output by the neural network model i The corresponding predicted probability, f(α i ) refers to the i-th activation function value.

5. The device for determining the posterior inclination angle of the tibial plateau according to claim 3, characterized in that: The neural network model includes 5 convolutional layers and 5 pooling layers, each convolutional layer contains two 3×3×3 convolution kernels; each pooling layer contains 2×2×2 pooling kernels.

6. The device for determining the posterior inclination angle of the tibial plateau according to claim 1, characterized in that: The calculation module calculates the Mahalanobis distances from all points in the target area to the rotation plane using the following formula: In the above formula, a j refers to the coordinate vector of the jth point in the target area, b j Refers to the coordinate vector of the projection point of the jth point in the target area on the rotation plane, D m (a j ) refers to the Mahalanobis distance from the jth point in the target area to the rotation plane, Σ -1 Refers to the inverse matrix of the covariance matrix of the target region.

7. The device for determining the posterior inclination angle of the tibial plateau according to claim 1, characterized in that: The preset angle is 0.5 degrees.

Images