Acetabular rotation center calculation method, device, equipment and readable storage medium
Through hip joint CT data processing and MLP neural network feature extraction, the rapid and accurate calculation of the acetabular rotation center is achieved, and the problems of large errors and low efficiency in traditional methods are solved, and a high-precision acetabular rotation center calculation method is provided.
Patent Information
- Application Number
- CN202310889643.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-19
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2043-07-19
AI Technical Summary
The traditional acetabular rotation center calculation method has large errors, low accuracy and low working efficiency, making it difficult to quickly and accurately calculate the acetabular rotation center.
By acquiring hip joint CT data, segmenting the femur and pelvis, positioning the femoral head position, collecting acetabular points, increasing the acquisition point, inputting the pre-trained acetabular center recognition network model, outputting the acetabular rotation center and radius, using MLP neural network for Feature extraction and global feature mapping, combined with maximum pooling operation, realize accurate calculation of the acetabular rotation center.
It significantly improves the calculation accuracy and speed of the acetabular rotation center, provides a fast and accurate calculation method, and solves the problems of large errors and low efficiency in traditional methods.
Smart Images

Figure CN116993824B_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of deep learning intelligent recognition technology, and in particular, relates to a method, device, equipment and computer-readable storage medium for calculating the acetabulum rotation center. Background Art
[0002] The traditional method for calculating the acetabulum rotation center is to mark it by professional physicians with many years of clinical experience, or to collect a certain number of points on the acetabulum and calculate it using traditional spherical fitting algorithms such as the least squares method, iterative closest point algorithm, and RANSAC algorithm. The calculation error is large, the accuracy is low, and the work efficiency is low.
[0003] Therefore, how to calculate the acetabulum rotation center quickly and accurately is a technical problem that technical personnel in this field need to solve urgently. Summary of the invention
[0004] The embodiments of the present application provide a method, device, equipment and computer-readable storage medium for calculating the rotation center of the acetabulum, which can calculate the rotation center of the acetabulum quickly and accurately.
[0005] In a first aspect, an embodiment of the present application provides a method for calculating an acetabulum rotation center, comprising:
[0006] Acquire hip joint CT data;
[0007] Segment the femur and pelvis of the hip joint CT data to obtain the femur segmentation results and pelvis results;
[0008] Locate the femoral head according to the femoral segmentation result;
[0009] Locate the acetabulum;
[0010] Acquisition of acetabular points;
[0011] Based on the acetabular position and acetabular point, increase the acetabular collection point;
[0012] The acetabulum collection points are input into a preset acetabulum center recognition network model, and the acetabulum rotation center and radius are output; wherein the acetabulum center recognition network model is obtained based on the acetabulum center recognition network through model training.
[0013] Optional, data collection, including:
[0014] Collect hip joint CT data, coordinates of acetabular point, coordinates and radius of acetabular rotation center.
[0015] Optionally, segment the femur and pelvis of the hip joint CT data, including:
[0016] The femur and pelvis of the hip joint CT data are segmented using a 2D segmentation algorithm or a 3D segmentation algorithm.
[0017] Optionally, the femoral head is located according to the femoral segmentation result, including:
[0018] The area 30 layers below the top of the femoral region is determined as the femoral head position.
[0019] Optionally, based on the acetabular position and acetabular points, additional acetabular collection points can be added, including:
[0020] For each collected point, take the point as the center of the sphere and search for points on the acetabulum within a radius of r = 5 mm;
[0021] When the number of points within the radius exceeds 100, it is downsampled to 100 points;
[0022] When the number of points within the radius is less than 100, the search radius is increased until the number of points within the search radius exceeds 100;
[0023] After adding the collection points, the number of point sets is 1000 points.
[0024] Optionally, in the acetabulum center recognition network, for the input N*3 data, the network first extracts features through Transform, maps it to a 64-dimensional space through an MLP structure, and finally maps it to a 1024-dimensional space;
[0025] Through the maximum pooling operation, only the largest one is retained in all 1024-dimensional channels. The resulting 1×1024 vector is the global feature of the N point clouds.
[0026] For regression problems, this global feature is directly passed through MLP to output the coordinates and radius of the acetabulum rotation center.
[0027] Optionally, the MLP structure includes three layers: input layer, hidden layer and output layer. The different layers of the MLP neural network are fully connected. The forward propagation of the MLP needs to add an activation function.
[0028] The weight update uses the gradient descent method: the weight is updated along the negative direction of the gradient to reduce the function value; the gradient refers to a vector, and the direction is the direction in which the directional derivative reaches the maximum value;
[0029] Learning rate: set to 0.0001;
[0030] The loss function uses the L2 regularization loss function.
[0031] In a second aspect, an embodiment of the present application provides a device for calculating the acetabulum rotation center, the device comprising:
[0032] A data acquisition module, used for acquiring hip joint CT data;
[0033] The bone segmentation module is used to segment the femur and pelvis of the hip joint CT data to obtain the femur segmentation results and pelvis results;
[0034] A femoral head position positioning module, used to locate the femoral head according to the femoral segmentation result;
[0035] An acetabulum position positioning module, used for positioning the acetabulum;
[0036] Acetabular point acquisition module, used for acquiring acetabular points;
[0037] Acetabular collection point expansion module, used to increase acetabular collection points based on acetabular position and acetabular point;
[0038] The acetabulum rotation center and radius module is used to input the acetabulum collection points into a preset acetabulum center recognition network model and output the acetabulum rotation center and radius; wherein the acetabulum center recognition network model is obtained based on the acetabulum center recognition network through model training.
[0039] In a third aspect, an embodiment of the present application provides an electronic device, the electronic device comprising: a processor and a memory storing computer program instructions;
[0040] When the processor executes the computer program instructions, the method for calculating the acetabulum rotation center as shown in the first aspect is implemented.
[0041] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having computer program instructions stored thereon, and when the computer program instructions are executed by a processor, the method for calculating the acetabulum rotation center as shown in the first aspect is implemented.
[0042] The acetabular rotation center calculation method, device, equipment and computer-readable storage medium of the embodiments of the present application can calculate the acetabular rotation center quickly and accurately.
[0043] The acetabulum rotation center calculation method comprises: acquiring hip joint CT data; segmenting the femur and pelvis of the hip joint CT data to obtain a femur segmentation result and a pelvis result; locating the femoral head position according to the femur segmentation result; locating the acetabulum position; collecting acetabulum points; adding acetabulum collection points based on the acetabulum position and acetabulum points; inputting the acetabulum collection points into a preset acetabulum center recognition network model, and outputting the acetabulum rotation center and radius; wherein the acetabulum center recognition network model is obtained through model training based on the acetabulum center recognition network. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0045] Figure 1 is a flow chart of a method for calculating the acetabulum rotation center provided by an embodiment of the present application;
[0046] Figure 2 is a flow chart of a method for calculating the acetabulum rotation center provided by an embodiment of the present application;
[0047] Figure 3 is a schematic diagram of the coordinates of the acetabulum point provided by an embodiment of the present application;
[0048] Figure 4 is a schematic diagram of increasing acetabular collection points provided by an embodiment of the present application;
[0049] Figure 5 is a schematic diagram of a network for identifying the acetabulum rotation center provided by an embodiment of the present application;
[0050] Figure 6 is a schematic structural diagram of a device for calculating the acetabulum rotation center provided by an embodiment of the present application;
[0051] Figure 7 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION
[0052] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.
[0053] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0054] The traditional method for calculating the acetabulum rotation center is to mark it by professional physicians with many years of clinical experience, or to collect a certain number of points on the acetabulum and calculate it using traditional spherical fitting algorithms such as the least squares method, iterative closest point algorithm, and RANSAC algorithm. The calculation error is large, the accuracy is low, and the work efficiency is low.
[0055] In order to solve the problems of the prior art, the embodiments of the present application provide a method, device, apparatus and computer-readable storage medium for calculating the rotation center of the acetabulum. The method for calculating the rotation center of the acetabulum provided in the embodiments of the present application is first introduced below.
[0056] Figure 1 FIG. 2 is a flow chart of a method for calculating the acetabulum rotation center provided by an embodiment of the present application. Figure 1 As shown, the method for calculating the acetabulum rotation center includes:
[0057] S101, acquiring hip joint CT data;
[0058] In one embodiment, data collection includes:
[0059] Collect hip joint CT data, coordinates of acetabular point, coordinates and radius of acetabular rotation center.
[0060] Among them, the schematic diagram of the coordinates of the acetabulum point is as follows Figure 3 shown.
[0061] S102, segmenting the femur and pelvis of the hip joint CT data to obtain a femur segmentation result and a pelvis segmentation result;
[0062] In one embodiment, segmenting the femur and pelvis of the hip joint CT data includes:
[0063] The femur and pelvis of the hip joint CT data are segmented using a 2D segmentation algorithm or a 3D segmentation algorithm.
[0064] Specifically, the hip joint data is segmented to obtain the pelvis and femur. The segmentation algorithm may use a 2D segmentation algorithm or a 3D segmentation algorithm, such as a 2D-unet or 3D-unet network.
[0065] S103, locating the femoral head according to the femoral segmentation result;
[0066] In one embodiment, locating the femoral head according to the femoral segmentation result includes:
[0067] The area 30 layers below the top of the femoral region is determined as the femoral head position.
[0068] Specifically, the position of the femoral head is determined according to the segmented femoral region, and the position of the femoral head is the region 30 layers below the top of the femoral region, where 30 layers is an empirical value obtained by analyzing multiple CT data.
[0069] S104, locating the acetabulum;
[0070] Specifically, according to the determined position of the femoral head area, the area in contact with the pelvis is selected, namely the acetabulum area.
[0071] S105, collecting acetabular points;
[0072] S106, based on the acetabulum position and acetabulum point, adding acetabulum collection points;
[0073] In one embodiment, based on the acetabular position and acetabular point, acetabular collection points are added, including:
[0074] For each collected point, take the point as the center of the sphere and search for points on the acetabulum within a radius of r = 5 mm;
[0075] When the number of points within the radius exceeds 100, it is downsampled to 100 points;
[0076] When the number of points within the radius is less than 100, the search radius is increased until the number of points within the search radius exceeds 100;
[0077] After adding the collection points, the number of point sets is 1000 points.
[0078] Specifically, in the data collection stage, a certain number of point sets are collected in the selected acetabulum area. The number of points collected here is relatively small (10 points), so the collected acetabulum point set needs to be expanded. The expansion method here is to use each collected point as the center of the sphere and search for points on the acetabulum within a radius of r = 5 mm. When the number of points within the radius exceeds 100 points, it is downsampled to 100 points. When the number of points within the radius is less than 100 points, the search radius is increased until the number of points within the search radius exceeds 100 points. Therefore, the number of point sets after increasing the collection points is 1000 points. After adding data points, a 3D data point cloud is formed. Among them, the schematic diagram of increasing acetabulum collection points is shown as follows. Figure 4 shown.
[0079] This method uses amplification to form a three-dimensional data point cloud from the few acetabulum points collected, solving the problem that the data is too small to be used for neural network training. Afterwards, it can be used to train the neural network through the point cloud data to obtain the center point coordinates and radius of the acetabulum.
[0080] S107, inputting the acetabulum collection points into a preset acetabulum center recognition network model, and outputting the acetabulum rotation center and radius; wherein the acetabulum center recognition network model is obtained through model training based on the acetabulum center recognition network.
[0081] In one embodiment, in the acetabulum center recognition network, for the input N*3 data, the network first extracts features through Transform, maps it to a 64-dimensional space through an MLP structure, and finally maps it to a 1024-dimensional space;
[0082] Through the maximum pooling operation, only the largest one is retained in all 1024-dimensional channels. The resulting 1×1024 vector is the global feature of the N point clouds.
[0083] For regression problems, this global feature is directly passed through MLP to output the coordinates and radius of the acetabulum rotation center.
[0084] Specifically, for the acetabular rotation center and radius recognition network, the structure is as follows: the input data is in N*3 three-dimensional coordinate point data format, where N represents the number of points 1000, and the output is [X, Y, Z, R], where X, Y, Z represent the coordinates of the acetabular rotation center, and R represents the acetabular radius. The network structure is as follows: Figure 5 shown.
[0085] In the acetabulum center recognition network, for the input N*3 data, the network first extracts features through Transform, maps it to a 64-dimensional space through the MLP structure, and finally maps it to a 1024-dimensional space. At this time, for each point, there is a 1024-dimensional vector representation, and such a vector representation is obviously redundant for a 3-dimensional point cloud. Therefore, the maximum pooling operation is introduced at this time to retain only the largest one in all 1024-dimensional channels. The 1×1024 vector obtained in this way is the global feature of the N point clouds. For regression problems, this global feature is directly output through MLP to output the coordinates and radius of the acetabulum rotation center.
[0086] In one embodiment, the MLP structure includes three layers: an input layer, a hidden layer, and an output layer. Different layers of the MLP neural network are fully connected. The forward propagation of the MLP needs to add an activation function. The weight update uses the gradient descent method: the weight is updated along the negative direction of the gradient to reduce the function value. The gradient refers to a vector, and the direction is the direction in which the directional derivative reaches the maximum value.
[0087] Learning rate: set to 0.0001;
[0088] The loss function uses the L2 regularization loss function.
[0089] Specifically, the MLP structure consists of three layers: input layer, hidden layer and output layer. The different layers of the MLP neural network are fully connected. The forward propagation of MLP requires the addition of an activation function, because if there is no activation function, the network will actually degenerate into a single-layer network.
[0090] The weight update uses the gradient descent method: the weight is updated in the negative direction of the gradient to reduce the function value. The gradient refers to a vector whose direction is the direction in which the directional derivative reaches the maximum value.
[0091] Learning Rate: Set to 0.0001.
[0092] The loss function uses the L2 regularization loss function:
[0093]
[0094] Figure 6 1 is a schematic diagram of the structure of a device for calculating the acetabulum rotation center provided by an embodiment of the present application, the device for calculating the acetabulum rotation center comprises:
[0095] Data acquisition module 601, used to acquire hip joint CT data;
[0096] The bone segmentation module 602 is used to segment the femur and pelvis of the hip joint CT data to obtain the femur segmentation result and the pelvis result;
[0097] A femoral head position positioning module 603 is used to locate the femoral head according to the femoral segmentation result;
[0098] Acetabular position positioning module 604, used to locate the acetabulum position;
[0099] Acetabular point acquisition module 605, used for acquiring acetabular points;
[0100] An acetabulum collection point expansion module 606 is used to increase acetabulum collection points based on the acetabulum position and acetabulum point;
[0101] The acetabulum rotation center and radius module 607 is used to input the acetabulum acquisition points into a preset acetabulum center recognition network model, and output the acetabulum rotation center and radius; wherein the acetabulum center recognition network model is obtained based on the acetabulum center recognition network through model training.
[0102] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown.
[0103] The electronic device may include a processor 701 and a memory 702 storing computer program instructions.
[0104] Specifically, the processor 701 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.
[0105] The memory 702 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 702 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 702 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 702 may be inside or outside the electronic device. In a particular embodiment, the memory 702 may be a non-volatile solid-state memory.
[0106] In one embodiment, the memory 702 may be a read-only memory (ROM). In one embodiment, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or a flash memory, or a combination of two or more of these.
[0107] The processor 701 reads and executes the computer program instructions stored in the memory 702 to implement any one of the methods for calculating the acetabulum rotation center in the above embodiments.
[0108] In one example, the electronic device may further include a communication interface 703 and a bus 710. Figure 7 As shown, the processor 701, the memory 702, and the communication interface 703 are connected via a bus 710 and communicate with each other.
[0109] The communication interface 703 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.
[0110] Bus 710 includes hardware, software or both, and the parts of electronic equipment are coupled to each other.For example, but not limitation, bus may include accelerated graphics port (AGP) or other graphics bus, enhanced industrial standard architecture (EISA) bus, front side bus (FSB), hypertransport (HT) interconnection, industrial standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 710 may include one or more buses. Although the present application embodiment describes and shows a specific bus, the application considers any suitable bus or interconnection.
[0111] In addition, in combination with the acetabular rotation center calculation method in the above embodiment, the embodiment of the present application can provide a computer-readable storage medium for implementation. The computer-readable storage medium stores computer program instructions; when the computer program instructions are executed by a processor, any one of the acetabular rotation center calculation methods in the above embodiment is implemented.
[0112] It should be clear that the present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.
[0113] The functional modules shown in the above-described block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. Programs or code segments can be stored in machine-readable media, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable media" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.
[0114] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be performed simultaneously.
[0115] The above reference is according to the method of the embodiment of the present application, the flow chart of the device (system) and the computer program product and / or the block diagram described various aspects of the present application.It should be understood that each square box in the flow chart and / or the block diagram and the combination of each square box in the flow chart and / or the block diagram can be realized by computer program instructions.These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine so that these instructions executed by the processor of the computer or other programmable data processing device enable the realization of the function / action specified in one or more square boxes of the flow chart and / or the block diagram.Such a processor can be but is not limited to a general-purpose processor, a special-purpose processor, a special application processor or a field programmable logic circuit.It can also be understood that each square box in the block diagram and / or the flow chart and the combination of the square boxes in the block diagram and / or the flow chart can also be realized by the dedicated hardware that performs the specified function or action, or can be realized by the combination of dedicated hardware and computer instructions.
[0116] The above is only a specific implementation of the present application. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present application is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the protection scope of this application.
Claims
1. A method for calculating the acetabulum rotation center, characterized in that: include: Acquire hip joint CT data; Segment the femur and pelvis of the hip joint CT data to obtain the femur segmentation results and pelvis results; Locate the femoral head according to the femoral segmentation result; Locate the acetabulum; Acquisition of acetabular points; Based on the acetabular position and acetabular point, increase the acetabular collection point; Input the acetabulum collection points into the preset acetabulum center recognition network model, and output the acetabulum rotation center and radius; wherein the acetabulum center recognition network model is obtained based on the acetabulum center recognition network through model training; Specifically, for the acetabular rotation center and radius recognition network, the structure is as follows: the input data is in N*3 three-dimensional coordinate point data format, where N represents the number of points 1000, and the output is [X, Y, Z, R], where X, Y, Z represent the coordinates of the acetabular rotation center, and R represents the acetabular radius; In the acetabulum center recognition network, for the input N*3 data, the network first extracts features through Transform, maps it to a 64-dimensional space through the MLP structure, and finally maps it to a 1024-dimensional space; at this time, for each point, there is a 1024-dimensional vector representation, and such a vector representation is obviously redundant for a 3-dimensional point cloud. Therefore, the maximum pooling operation is introduced at this time to retain only the largest one in all 1024-dimensional channels. The 1×1024 vector obtained in this way is the global feature of the N point clouds; for the regression problem, this global feature is directly output through MLP to output the coordinates and radius of the acetabulum rotation center; The MLP structure consists of three layers: input layer, hidden layer and output layer. The different layers of the MLP neural network are fully connected. The forward propagation of the MLP requires the addition of an activation function. The weight update uses the gradient descent method: the weight is updated along the negative direction of the gradient to reduce the function value; the gradient refers to a vector, and the direction is the direction in which the directional derivative reaches the maximum value; Learning rate: set to 0.0001; The loss function uses the L2 regularization loss function; Data collection, including: Collect hip joint CT data, coordinates of acetabular point, coordinates and radius of acetabular rotation center; Segmentation of the femur and pelvis of the hip joint CT data, including: Use 2D segmentation algorithm or 3D segmentation algorithm to segment the femur and pelvis of hip joint CT data; Locate the femoral head according to the femoral segmentation results, including: The area 30 layers below the top of the femoral region is determined as the femoral head position; Based on the acetabular position and acetabular point, additional acetabular collection points are added, including: For each collected point, take the point as the center of the sphere and search for points on the acetabulum within a radius of r = 5 mm; When the number of points within the radius exceeds 100, it is downsampled to 100 points; When the number of points within the radius is less than 100, the search radius is increased until the number of points within the search radius exceeds 100; After increasing the number of collected points, the number of point sets is 1000 points to form a three-dimensional data point cloud; the three-dimensional data point cloud is formed by amplification, and the acetabulum rotation center and radius are obtained through neural network training of the three-dimensional data point cloud.
2. A device for calculating the acetabulum rotation center, characterized in that: The device comprises: A data acquisition module, used for acquiring hip joint CT data; The bone segmentation module is used to segment the femur and pelvis of the hip joint CT data to obtain the femur segmentation results and pelvis results; A femoral head position positioning module, used to locate the femoral head according to the femoral segmentation result; An acetabulum position positioning module, used for positioning the acetabulum; Acetabular point acquisition module, used for acquiring acetabular points; Acetabular collection point expansion module, used to increase acetabular collection points based on acetabular position and acetabular point; Acetabular rotation center and radius module, used to input acetabular collection points into a preset acetabular center recognition network model, and output the acetabular rotation center and radius; wherein the acetabular center recognition network model is obtained through model training based on the acetabular center recognition network; Specifically, for the acetabular rotation center and radius recognition network, the structure is as follows: the input data is in N*3 three-dimensional coordinate point data format, where N represents the number of points 1000, and the output is [X, Y, Z, R], where X, Y, Z represent the coordinates of the acetabular rotation center, and R represents the acetabular radius; In the acetabulum center recognition network, for the input N*3 data, the network first extracts features through Transform, maps it to a 64-dimensional space through the MLP structure, and finally maps it to a 1024-dimensional space; at this time, for each point, there is a 1024-dimensional vector representation, and such a vector representation is obviously redundant for a 3-dimensional point cloud. Therefore, the maximum pooling operation is introduced at this time to retain only the largest one in all 1024-dimensional channels. The 1×1024 vector obtained in this way is the global feature of the N point clouds; for the regression problem, this global feature is directly output through MLP to output the coordinates and radius of the acetabulum rotation center; The MLP structure consists of three layers: input layer, hidden layer and output layer. The different layers of the MLP neural network are fully connected. The forward propagation of the MLP requires the addition of an activation function. The weight update uses the gradient descent method: the weight is updated along the negative direction of the gradient to reduce the function value; the gradient refers to a vector, and the direction is the direction in which the directional derivative reaches the maximum value; Learning rate: set to 0.0001; The loss function uses the L2 regularization loss function; Data collection, including: Collect hip joint CT data, coordinates of acetabular point, coordinates and radius of acetabular rotation center; Segmentation of the femur and pelvis of the hip joint CT data, including: Use 2D segmentation algorithm or 3D segmentation algorithm to segment the femur and pelvis of hip joint CT data; Locate the femoral head according to the femoral segmentation results, including: The area 30 layers below the top of the femoral region is determined as the femoral head position; Based on the acetabular position and acetabular point, additional acetabular collection points are added, including: For each collected point, take the point as the center of the sphere and search for points on the acetabulum within a radius of r = 5 mm; When the number of points within the radius exceeds 100, it is downsampled to 100 points; When the number of points within the radius is less than 100, the search radius is increased until the number of points within the search radius exceeds 100; After increasing the number of collected points, the number of point sets is 1000 points to form a three-dimensional data point cloud; the three-dimensional data point cloud is formed by amplification, and the acetabulum rotation center and radius are obtained through neural network training of the three-dimensional data point cloud.
3. An electronic device, characterized in that: The electronic device comprises: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the method for calculating the acetabulum rotation center as claimed in claim 1 is implemented.
4. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the method for calculating the acetabulum rotation center according to claim 1 is implemented.
Citation Information
Patent Citations
Children femoral head rotation center detection method
CN111429417A
Total hip replacement preoperative planning method and device based on deep learning and X rays
CN111888059A
Hip joint rotation center detection method and imaging method under multi-attitude condition
CN113298783A
Method and system for recognizing force line by using neural network, storage medium and electronic equipment
CN113870261A
Acetabular cup position adjusting method and device based on reinforcement learning and storage medium
CN113962927A
