Method and device for judging abnormality of prosthetic external rotation angle based on knee cross-sectional ct

Abstract

The application provides a prosthesis external rotation angle abnormality judgment method and device based on knee joint cross-sectional CT, and the method comprises the following steps: acquiring a knee joint cross-sectional CT image; identifying knee joint bony landmark points based on a deep convolution model; calculating a prosthesis external rotation angle based on the knee joint bony landmark points; and judging whether the prosthesis external rotation angle is abnormal based on a preset angle range. In the application, the key bony landmark points of the prosthesis are automatically identified based on the CT cross-sectional image, and the rotation angle is fitted through an algorithm, so that the identification accuracy is greatly improved.

Description

Technical Field

[0001] This application relates to the field of medical image processing technology, and more specifically, to a method and device for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint. Background Technology

[0002] Assessing the external rotation angle of the prosthesis after knee arthroplasty is a key indicator for evaluating postoperative function and preventing complications such as patellar instability and joint pain. Currently, the assessment of the external rotation angle of the prosthesis is mainly conducted through intraoperative mechanical positioning and imaging.

[0003] However, the accuracy of the tools used in the former procedure is easily affected by the operator's experience, while the latter relies on manual identification of bony landmarks, resulting in low measurement accuracy. Summary of the Invention

[0004] The problem addressed by this application is the low accuracy of assessing the external rotation angle of the prosthesis.

[0005] To address the aforementioned issues, the first aspect of this application provides a method for judging abnormal external rotation angles of the prosthesis based on transverse CT scans of the knee joint, comprising:

[0006] Obtain cross-sectional CT images of the knee joint;

[0007] Identify bony landmarks of the knee joint based on a deep convolution model;

[0008] Calculate the external rotation angle of the prosthesis based on the aforementioned bony landmarks of the knee joint;

[0009] Based on a preset angle range, determine whether the external rotation angle of the prosthesis is abnormal.

[0010] The second aspect of this application provides a manufacturing system for a method of judging abnormal external rotation angle of a prosthesis based on transverse CT of the knee joint, comprising:

[0011] Image acquisition module, which is used to acquire cross-sectional CT images of the knee joint;

[0012] The landmark recognition module is used to identify bony landmarks of the knee joint based on a deep convolution model.

[0013] An external rotation calculation module is used to calculate the external rotation angle of the prosthesis based on the bony landmarks of the knee joint;

[0014] The anomaly detection module is used to determine whether the external rotation angle of the prosthesis is abnormal based on a preset angle range.

[0015] A third aspect of this application provides an electronic device, including: a memory and a processor; the memory being configurable to store a program, and the processor being coupled to the memory for executing the program in the memory for:

[0016] Obtain cross-sectional CT images of the knee joint;

[0017] Identify bony landmarks of the knee joint based on a deep convolution model;

[0018] Calculate the external rotation angle of the prosthesis based on the aforementioned bony landmarks of the knee joint;

[0019] Based on a preset angle range, determine whether the external rotation angle of the prosthesis is abnormal.

[0020] The fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the aforementioned method for judging abnormal external rotation angle of a prosthesis based on transverse CT of the knee joint.

[0021] In this application, key bony landmarks of the prosthesis are automatically identified based on CT cross-sectional images, and the rotation angle is fitted by an algorithm, thereby greatly improving the accuracy of identification. Attached Figure Description

[0022] Figure 1 This is a flowchart of a method for judging abnormal external rotation angle of a prosthesis based on transverse CT of the knee joint according to an embodiment of this application;

[0023] Figure 2 This is a model architecture diagram of the method for judging abnormal external rotation angle of prosthesis based on cross-sectional CT of knee joint according to an embodiment of this application;

[0024] Figure 3 This is an architecture diagram of the clustering filtering module of the method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint according to an embodiment of this application;

[0025] Figure 4 This is a schematic diagram of a device for judging abnormal external rotation angle of a prosthesis based on cross-sectional CT of the knee joint according to an embodiment of this application.

[0026] Figure 5 This is an architectural diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0027] To make the above-mentioned objects, features, and advantages of this application more apparent and understandable, specific embodiments of this application will be described in detail below with reference to the accompanying drawings. Although exemplary embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0028] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application shall have the ordinary meaning as understood by one of ordinary skill in the art to which this application pertains.

[0029] Assessing the external rotation angle of the prosthesis after knee arthroplasty is a key indicator for evaluating postoperative function and preventing complications such as patellar instability and joint pain. Current assessment methods include:

[0030] Intraoperative mechanical positioning: The rotation angle of the prosthesis is adjusted using intraoperative mechanical guiding tools.

[0031] Disadvantages: The precision of surgical instruments may be affected by soft tissue tension and operator experience, making it difficult to reflect the postoperative condition.

[0032] Imaging assessment:

[0033] CT scan: The posterior condyle line of the femoral prosthesis and the axis of rotation of the tibial prosthesis are manually measured using CT cross-sectional images.

[0034] Disadvantages: Traditional measurement methods rely on manual identification of bony landmarks, and the results are greatly affected by the operator's experience, resulting in low measurement efficiency and poor consistency.

[0035] MRI scan: used to assess the relationship between soft tissue and prosthesis.

[0036] Disadvantages: MRI has difficulty clearly displaying bony landmarks around the metal prosthesis, and it is expensive and complicated to operate.

[0037] This application provides the above-described method for judging prosthetic external rotation angle abnormalities based on knee joint transverse CT scans. The specific scheme of this method is described in [the following text is missing from the original]. Figures 1-4 As shown, this method can be executed by a device for judging abnormal external rotation angle of the prosthesis based on transverse CT of the knee joint. This device can be integrated into electronic devices such as computers, servers, computer clusters, and data centers. Combined with... Figure 1 As shown, the method for judging abnormal prosthesis external rotation angle based on knee joint transverse CT includes:

[0038] S101, acquire cross-sectional CT images of the knee joint;

[0039] In this application, the knee joint transverse CT image is an image taken on the horizontal plane of the knee joint using computed tomography (CT) technology, showing the cross-sectional structure inside the knee joint.

[0040] S102, based on a deep convolution model, identifies bony landmarks of the knee joint;

[0041] S103, Calculate the external rotation angle of the prosthesis based on the aforementioned bony landmarks of the knee joint;

[0042] S104, based on a preset angle range, determines whether the external rotation angle of the prosthesis is abnormal.

[0043] In this application, key bony landmarks of the prosthesis are automatically identified based on CT cross-sectional images, and the rotation angle is fitted by an algorithm, thereby greatly improving the accuracy of identification.

[0044] In this application, the calculated external rotation angle of the femoral prosthesis and the rotation angle of the tibial prosthesis are compared with preset normal ranges (e.g., the external rotation angle of the femur is typically 3°-5°, and the rotation angle of the tibia is 0°-5°):

[0045] If the angle exceeds the preset range, the external rotation angle of the prosthesis is judged to be abnormal; if the angle is within the preset range, the external rotation angle of the prosthesis is judged to be normal.

[0046] The existing technology has the following problems:

[0047] Human error: Existing CT measurement methods rely on doctors manually identifying anatomical landmarks (such as the posterior condyle line of the femur), which has significant subjectivity and measurement error.

[0048] Poor consistency of results: Due to differences in the experience and judgment criteria of different operators, the measurement results of the same image may vary significantly, affecting clinical decision-making.

[0049] Inefficient: Manual measurement requires identifying anatomical landmarks in CT images one by one, which is labor-intensive and time-consuming, making it difficult to meet large-scale clinical needs.

[0050] In this application, based on CT cross-sectional images, key bony landmarks of the prosthesis (such as the posterior condyle line of the femur and the prosthesis plateau of the tibia) are automatically identified, and the rotation angle is fitted by an algorithm.

[0051] This application employs a deep learning model to reduce manual intervention and provide automated, standardized external rotation angle measurement.

[0052] In this application, by combining anatomical and prosthesis geometric features, a quantitative analysis of abnormal external rotation angles is achieved, and a visual diagnostic report is output.

[0053] In one embodiment, the knee joint transverse CT image covers the bone tissue at the posterior condyle of the femoral prosthesis and the bone tissue at the tibial prosthesis plateau.

[0054] In this application, the bone tissue at the posterior condyle of the femoral prosthesis is shown: the prosthesis structure of the posterior condyle of the femur is displayed to fit the posterior condyle line of the femoral prosthesis.

[0055] In this application, the bone tissue at the tibial prosthesis platform is shown: the center of the tibial prosthesis platform is displayed to fit the axis of rotation of the tibial prosthesis.

[0056] In one embodiment, the bony landmarks include the femoral prosthesis posterior condyle key point and the tibial prosthesis plateau center point.

[0057] In this application, the key point of the posterior condyle of the femoral prosthesis is: the prosthesis feature point in the posterior condyle region of the femur.

[0058] In this application, the center point of the tibial prosthesis platform is defined as the center of the tibial prosthesis platform.

[0059] In one embodiment, calculating the external rotation angle of the prosthesis based on the bony landmarks of the knee joint includes:

[0060] Based on the aforementioned bony landmarks of the knee joint, the femoral anatomical axis and the prosthesis posterior condyle line, the tibial anatomical axis and the prosthesis rotation axis were fitted.

[0061] Calculate the angle between the posterior condyle line of the femoral prosthesis and the anterior-posterior midline of the femur, and use it as the external rotation angle of the femoral prosthesis;

[0062] Calculate the angle between the rotation axis of the tibial prosthesis and the anatomical axis of the tibia, and use it as the rotation angle of the tibial prosthesis.

[0063] In this application, the femoral anatomical axis is defined as the natural long axis of the femur fitted using anatomical landmarks.

[0064] In this application, femoral anatomical landmarks (bone landmarks of the knee joint) include anatomical features of the proximal femur (e.g., the center of the femoral head or neck) and the distal femur (e.g., the center point of the intercondylar line of the femur).

[0065] In this application, the posterior condyle line of the prosthesis is obtained by fitting key points of the posterior condyle of the femoral prosthesis.

[0066] In this application, the key points of the posterior condyle of the femoral prosthesis (bone landmarks of the knee joint) are two landmarks (two ends of the posterior condyle) at the posterior condyle of the femoral prosthesis, and the two are fitted with a straight line.

[0067] In this application, the anatomical axis of the tibia is determined by fitting the natural long axis of the tibia to bony landmarks.

[0068] In this application, the tibial anatomical axis is fitted by bony landmarks at the proximal and distal ends of the tibia (the center of the tibial plateau and the center of the tibial condyle). The bony landmarks of the knee joint include the center of the tibial plateau and the center of the tibial condyle.

[0069] In this application, the prosthesis rotation axis is determined by fitting the rotation axis through the center point of the tibial prosthesis platform and the prosthesis connection point. The prosthesis rotation axis is the axis of rotation of the tibial prosthesis.

[0070] In this application, the bony landmarks of the knee joint include: the center point of the tibial prosthesis platform and the prosthesis connection point.

[0071] In this application, the angle between the posterior condyle line of the femoral prosthesis and the anterior-posterior midline of the femur (i.e., the anatomical axis) is the external rotation angle of the femoral prosthesis. This indicates whether the direction of rotation of the femoral prosthesis is aligned with the anterior-posterior midline of the femur.

[0072] In this application, the angle between the rotation axis of the tibial prosthesis and the anatomical axis of the tibia is the rotation angle of the tibial prosthesis. This indicates whether the rotation direction of the tibial prosthesis is aligned with the natural axis of the tibia.

[0073] In one embodiment, after acquiring the transverse CT image of the knee joint, the method further includes:

[0074] The knee joint transverse CT image is calibrated to obtain a calibrated knee joint transverse CT image.

[0075] In this application, knee joint CT images are calibrated to eliminate errors caused by angle deviations or equipment problems during the shooting process, ensuring image standardization.

[0076] In this application, the calibration includes: adjusting the knee joint center alignment to the reference frame of the CT image; and ensuring that the anatomical positions of the femoral prosthesis posterior condyle and tibial prosthesis platform are not affected by device offset. The calibrated images will provide accurate input data to reduce angle calculation errors.

[0077] In this application, image calibration is used to adjust scanning parameters to reduce artifact interference.

[0078] In one implementation, combined with Figure 2 As shown, the identification of bony landmarks of the knee joint based on the depth convolution model includes:

[0079] Feature extraction was performed on the transverse CT image of the knee joint to obtain the feature extraction map;

[0080] The feature extraction map is subjected to sequential clustering filtering to obtain the filtered feature map;

[0081] The filtered feature map is processed by a fully connected layer to obtain bony landmarks of the knee joint.

[0082] In this application, multi-layer convolution processing is performed on knee joint CT images to extract key features of the bony structure and generate a feature extraction map.

[0083] In this application, the multi-layer convolutional processing includes multiple convolutional layers: extracting local features (such as edges, textures, etc.) through convolutional kernels; pooling layers: reducing the size of feature maps through max pooling or average pooling to improve the computational efficiency of the model; activation functions: such as the ReLU function, used to introduce non-linear characteristics.

[0084] In this application, after multi-layer convolution and pooling operations, the obtained feature extraction map contains a high-dimensional feature representation of the knee joint cross-sectional CT image, preserving the bony structure information in the image.

[0085] In this application, the filtered feature map is passed through a fully connected layer to output specific knee joint bony landmarks, including: femoral prosthesis posterior condyle key point: locating the prosthesis feature point in the femoral posterior condyle region; tibial prosthesis platform center point: locating the center of the tibial prosthesis platform.

[0086] In this application, a fully connected network is used to transform the filtered feature map into specific skeletal landmark coordinates, and a label is assigned to each landmark.

[0087] In one implementation, combined with Figure 3 As shown, the clustering filtering process includes:

[0088] The input feature map is clustered to obtain a clustered feature map;

[0089] Position encoding is performed on the clustered feature map to obtain attention key values;

[0090] The attention key values ​​are subjected to frequency filtering, feedforward network processing, and superposition processing to obtain a superimposed feature map.

[0091] The restored feature map is obtained by performing restoration processing based on the input feature map and the superimposed feature map;

[0092] The recovery process is performed based on the recovered feature map and the superimposed feature map to obtain the output feature map.

[0093] The input feature map is the feature map output by the deep convolutional model, which is usually a high-dimensional feature matrix.

[0094] In this application, the clustering algorithm captures high- and low-frequency features in the input feature map to obtain favorable frequency components and generate cluster centers containing information about the differences between each category, which is the cluster feature map.

[0095] In this application, the clustering algorithm can be K-Means clustering, density clustering (DBSCAN), or hierarchical clustering.

[0096] In this application, location information is introduced into each cluster region through location encoding, thereby enhancing the model's ability to perceive spatial features.

[0097] In this application, location encoding is added to each cluster region of the cluster feature map to preserve spatial location information.

[0098] In this application, attention keys are generated, and each cluster region contains not only feature information but also spatial relationships.

[0099] In this application, the features are decomposed in the frequency domain to extract high-frequency (edge ​​information) and low-frequency (overall structure) features. Specifically, high-pass filtering extracts detailed features, such as edge and texture information; low-pass filtering extracts overall features, such as smoothing features of large areas.

[0100] In this application, the feedforward network enhances the complex relationships between features through nonlinear transformations. The input is a frequency-filtered feature map, which is processed through one or more fully connected networks; activation functions (such as ReLU) are used to introduce nonlinear characteristics.

[0101] In this application, features processed by frequency filtering and feedforward networks are superimposed to integrate multiple types of information. Pixel-by-pixel addition or stitching operations are used to superimpose features from different sources. The superimposed feature map contains both frequency domain and network-processed information.

[0102] In this application, feature recovery is performed by combining the original input feature map and the superimposed feature map, preserving the original information while enhancing the processed features. Specifically, this may include: feature fusion: performing pixel-by-pixel addition or other fusion methods on the input feature map and the superimposed feature map; nonlinear activation: using an activation function (such as ReLU) to process the fused features and enhance nonlinear characteristics.

[0103] In this application, a final high-quality output feature map is generated by further fusing the restored feature map and the superimposed feature map. Specifically, this may include: bidirectional fusion: further fusing the restored feature map and the superimposed feature map to fully utilize the information from both; generating the final feature map: compressing the size of the feature map using convolution or pooling operations, removing redundant information, and retaining key features; output: the final output feature map, which contains a high-dimensional feature representation of the bony landmarks of the knee joint.

[0104] In this application, AI algorithms are used to automatically analyze CT images, avoiding human error and making the measurement results more objective and consistent.

[0105] In this application, automated processing significantly reduces measurement time and improves clinical work efficiency.

[0106] The unified method for identifying anatomical landmarks and calculating angles in this application helps to standardize the diagnostic process for abnormal external rotation angles.

[0107] This application also includes: generating a graphical report containing landmarks, angle measurement results, and diagnostic conclusions for doctors' reference.

[0108] In this application, the measurement accuracy of the device was verified using real patient data, and the results showed that:

[0109] Compared with manual measurement, automated measurement has an error of less than ±1°, demonstrating good accuracy and repeatability.

[0110] The measurement time has been reduced from 15 minutes using traditional methods to approximately 2 minutes, significantly improving efficiency.

[0111] This application provides a device for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint, which is used to execute the method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint described above. The following is a detailed description of the device for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint.

[0112] like Figure 4 As shown, the device for judging abnormal external rotation angle of the prosthesis based on transverse CT of the knee joint includes:

[0113] Image acquisition module 101 is used to acquire cross-sectional CT images of the knee joint;

[0114] Landmark recognition module 102 is used to identify bony landmarks of the knee joint based on a deep convolution model;

[0115] External rotation calculation module 103 is used to calculate the external rotation angle of the prosthesis based on the bony landmarks of the knee joint;

[0116] The anomaly detection module 104 is used to determine whether the external rotation angle of the prosthesis is abnormal based on a preset angle range.

[0117] In one embodiment, the knee joint transverse CT image covers the bone tissue at the posterior condyle of the femoral prosthesis and the bone tissue at the tibial prosthesis plateau.

[0118] In one embodiment, the bony landmarks include the femoral prosthesis posterior condyle key point and the tibial prosthesis plateau center point.

[0119] In one embodiment, the external rotation calculation module 103 is further configured to:

[0120] Based on the aforementioned bony landmarks of the knee joint, the femoral anatomical axis and the prosthesis posterior condyle line, the tibial anatomical axis and the prosthesis rotation axis are fitted; the angle between the femoral prosthesis posterior condyle line and the anterior-posterior midline of the femur is calculated as the external rotation angle of the femoral prosthesis; the angle between the tibial prosthesis rotation axis and the tibial anatomical axis is calculated as the tibial prosthesis rotation angle.

[0121] In one embodiment, after acquiring the knee joint transverse CT image, the method further includes: calibrating the knee joint transverse CT image to obtain a calibrated knee joint transverse CT image.

[0122] In one embodiment, the marker recognition module 102 is further configured to:

[0123] Feature extraction was performed on the transverse CT image of the knee joint to obtain a feature extraction map; sequential clustering filtering was performed on the feature extraction map to obtain a filtered feature map; and fully connected processing was performed on the filtered feature map to obtain the bony landmarks of the knee joint.

[0124] In one embodiment, the marker recognition module 102 is further configured to:

[0125] Clustering is performed on the input feature map to obtain a clustered feature map; position encoding is performed on the clustered feature map to obtain attention key values; frequency filtering, feedforward network processing, and superposition processing are performed on the attention key values ​​to obtain a superimposed feature map; recovery processing is performed based on the input feature map and the superimposed feature map to obtain a recovered feature map; recovery processing is performed based on the recovered feature map and the superimposed feature map to obtain an output feature map.

[0126] The device for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint provided in the above embodiments of this application corresponds to the method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint provided in the embodiments of this application. Therefore, the specific content in this system corresponds to the method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint. The specific content can be referred to the records in the method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint. This application will not repeat the details.

[0127] The device for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint provided in the above embodiments of this application and the method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods used, run or implemented by the application stored therein.

[0128] The above describes the internal function and structure of a device for judging abnormal external rotation angle of a prosthesis based on transverse CT of the knee joint, such as... Figure 5 As shown, in practice, the device for judging abnormal external rotation angle of the prosthesis based on transverse CT of the knee joint can be implemented as an electronic device, including: memory 301 and processor 303.

[0129] Memory 301 can be configured to store a program.

[0130] Additionally, memory 301 can also be configured to store various other data to support operation on the electronic device. Examples of this data include instructions for any application or method used to operate on the electronic device, contact data, phonebook data, messages, pictures, videos, etc.

[0131] Memory 301 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. Processor 303, coupled to memory 301, is used to execute programs in memory 301 for:

[0132] Obtain cross-sectional CT images of the knee joint;

[0133] Identify bony landmarks of the knee joint based on a deep convolution model;

[0134] Calculate the external rotation angle of the prosthesis based on the aforementioned bony landmarks of the knee joint;

[0135] Based on a preset angle range, determine whether the external rotation angle of the prosthesis is abnormal.

[0136] In one embodiment, the knee joint transverse CT image covers the bone tissue at the posterior condyle of the femoral prosthesis and the bone tissue at the tibial prosthesis plateau.

[0137] In one embodiment, the bony landmarks include the femoral prosthesis posterior condyle key point and the tibial prosthesis plateau center point.

[0138] In one implementation, the processor 303 is further configured to:

[0139] Based on the aforementioned bony landmarks of the knee joint, the femoral anatomical axis and the prosthesis posterior condyle line, the tibial anatomical axis and the prosthesis rotation axis are fitted; the angle between the femoral prosthesis posterior condyle line and the anterior-posterior midline of the femur is calculated as the external rotation angle of the femoral prosthesis; the angle between the tibial prosthesis rotation axis and the tibial anatomical axis is calculated as the tibial prosthesis rotation angle.

[0140] In one embodiment, after acquiring the knee joint transverse CT image, the method further includes: calibrating the knee joint transverse CT image to obtain a calibrated knee joint transverse CT image.

[0141] In one implementation, the processor 303 is further configured to:

[0142] Feature extraction was performed on the transverse CT image of the knee joint to obtain a feature extraction map; sequential clustering filtering was performed on the feature extraction map to obtain a filtered feature map; and fully connected processing was performed on the filtered feature map to obtain the bony landmarks of the knee joint.

[0143] In one implementation, the processor 303 is further configured to:

[0144] Clustering is performed on the input feature map to obtain a clustered feature map; position encoding is performed on the clustered feature map to obtain attention key values; frequency filtering, feedforward network processing, and superposition processing are performed on the attention key values ​​to obtain a superimposed feature map; recovery processing is performed based on the input feature map and the superimposed feature map to obtain a recovered feature map; recovery processing is performed based on the recovered feature map and the superimposed feature map to obtain an output feature map.

[0145] In this application, Figure 5 The diagram only shows some components and does not mean that the electronic device includes only these components. Figure 5 The components shown.

[0146] The electronic device provided in this embodiment is based on the same inventive concept as the method for judging abnormal external rotation angle of the prosthesis based on cross-sectional CT of the knee joint provided in this application embodiment, and has the same beneficial effects as the methods adopted, run or implemented by the application stored therein.

[0147] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0148] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0149] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0150] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory. Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0151] This application also provides a computer-readable storage medium corresponding to the method for judging abnormal external rotation angle of prosthesis based on cross-sectional CT of knee joint provided in the foregoing embodiments, wherein a computer program (i.e., program product) is stored thereon. When the computer program is run by a processor, it executes the interactive image analysis assistance method for 3D aerial imaging provided in any of the foregoing embodiments.

[0152] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0153] The computer-readable storage medium provided in the above embodiments of this application and the interactive image analysis assistance method for 3D aerial imaging provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the applications stored therein.

[0154] It should be noted that numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known structures and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0155] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0156] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for judging abnormal external rotation angle of the prosthesis based on transverse CT of the knee joint, characterized in that, include: Obtain cross-sectional CT images of the knee joint; Identify bony landmarks of the knee joint based on a deep convolution model; Calculate the external rotation angle of the prosthesis based on the aforementioned bony landmarks of the knee joint; Based on a preset angle range, determine whether the external rotation angle of the prosthesis is abnormal; The method for identifying bony landmarks of the knee joint based on a depthwise convolution model includes: Feature extraction was performed on the transverse CT image of the knee joint to obtain the feature extraction map; The feature extraction map is subjected to sequential clustering filtering to obtain the filtered feature map; The filtered feature map is processed by a fully connected layer to obtain bony landmarks of the knee joint. The clustering filtering process includes: The input feature map is clustered to obtain a clustered feature map; Position encoding is performed on the clustered feature map to obtain attention key values; The attention key values ​​are subjected to frequency filtering, feedforward network processing, and superposition processing to obtain a superimposed feature map. The restored feature map is obtained by performing restoration processing based on the input feature map and the superimposed feature map; The recovery process is performed based on the recovered feature map and the superimposed feature map to obtain the output feature map.

2. The method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint according to claim 1, characterized in that, The CT images of the knee joint cross section cover the bone tissue at the posterior condyle of the femoral prosthesis and the bone tissue at the tibial prosthesis plateau.

3. The method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint according to claim 1, characterized in that, The bony landmarks include the femoral prosthesis posterior condyle key point and the tibial prosthesis plateau center point.

4. The method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint according to claim 3, characterized in that, The calculation of the prosthesis external rotation angle based on the bony landmarks of the knee joint includes: Based on the aforementioned bony landmarks of the knee joint, the femoral anatomical axis and the prosthesis posterior condyle line, the tibial anatomical axis and the prosthesis rotation axis were fitted. Calculate the angle between the posterior condyle line of the femoral prosthesis and the anterior-posterior midline of the femur, and use it as the external rotation angle of the femoral prosthesis; Calculate the angle between the rotation axis of the tibial prosthesis and the anatomical axis of the tibia, and use it as the rotation angle of the tibial prosthesis.

5. The method for judging abnormal external rotation angle of prosthesis based on transverse CT of knee joint according to claim 1, characterized in that, After acquiring the transverse CT image of the knee joint, the process also includes: The knee joint transverse CT image is calibrated to obtain a calibrated knee joint transverse CT image.

6. A device for judging abnormal external rotation angle of a prosthesis based on transverse CT of the knee joint, characterized in that, include: Image acquisition module, which is used to acquire cross-sectional CT images of the knee joint; The landmark recognition module is used to identify bony landmarks of the knee joint based on a deep convolution model. An external rotation calculation module is used to calculate the external rotation angle of the prosthesis based on the bony landmarks of the knee joint; The anomaly detection module is used to determine whether the external rotation angle of the prosthesis is abnormal based on a preset angle range. The method for identifying bony landmarks of the knee joint based on a depthwise convolution model includes: Feature extraction was performed on the transverse CT image of the knee joint to obtain the feature extraction map; The feature extraction map is subjected to sequential clustering filtering to obtain the filtered feature map; The filtered feature map is processed by a fully connected layer to obtain bony landmarks of the knee joint. The clustering filtering process includes: The input feature map is clustered to obtain a clustered feature map; Position encoding is performed on the clustered feature map to obtain attention key values; The attention key values ​​are subjected to frequency filtering, feedforward network processing, and superposition processing to obtain a superimposed feature map. The restored feature map is obtained by performing restoration processing based on the input feature map and the superimposed feature map; The recovery process is performed based on the recovered feature map and the superimposed feature map to obtain the output feature map.

7. An electronic device, characterized in that, include: Memory and processor; The memory is used to store programs; The processor, coupled to the memory, is used to execute the program for: Obtain cross-sectional CT images of the knee joint; Identify bony landmarks of the knee joint based on a deep convolution model; Calculate the external rotation angle of the prosthesis based on the aforementioned bony landmarks of the knee joint; Based on a preset angle range, determine whether the external rotation angle of the prosthesis is abnormal; The method for identifying bony landmarks of the knee joint based on a depthwise convolution model includes: Feature extraction was performed on the transverse CT image of the knee joint to obtain the feature extraction map; The feature extraction map is subjected to sequential clustering filtering to obtain the filtered feature map; The filtered feature map is processed by a fully connected layer to obtain bony landmarks of the knee joint. The clustering filtering process includes: The input feature map is clustered to obtain a clustered feature map; Position encoding is performed on the clustered feature map to obtain attention key values; The attention key values ​​are subjected to frequency filtering, feedforward network processing, and superposition processing to obtain a superimposed feature map. The restored feature map is obtained by performing restoration processing based on the input feature map and the superimposed feature map; The recovery process is performed based on the recovered feature map and the superimposed feature map to obtain the output feature map.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the method for judging abnormal external rotation angle of the prosthesis based on cross-sectional CT of the knee joint as described in any one of claims 1-5.

Images