A method, device, equipment, medium and product for evaluating postoperative bone transplant materials

Abstract

The present application discloses a method, device, equipment, medium and product for evaluating postoperative bone transplant materials, and relates to the field of bone transplant material evaluation. The method comprises: acquiring historical CBCT images at different time nodes, marking the historical CBCT images, and determining the CBCT images with the markings; based on the CBCT images with the markings at different time nodes, taking the historical CBCT images as input and the marked bone transplant material segmentation results as output, constructing and training a neural network model, and determining the trained neural network model; inputting the CBCT images to be tested at different time nodes into the trained neural network, and outputting the bone transplant material segmentation results; and evaluating the degree of change of the bone transplant material in the CBCT images to be tested according to the bone transplant material segmentation results corresponding to the different time nodes. The present application can reduce segmentation errors and labor costs, and improve segmentation efficiency.

Description

Technical Field

[0001] The present application relates to the field of bone transplant material evaluation, and in particular to a method, device, equipment, medium and product for evaluating postoperative bone transplant materials. Background Art

[0002] After maxillary sinus lift and implant surgery, accurate assessment of bone graft material changes is crucial to predict long-term implant survival. Traditional assessment methods mainly rely on manual marking and segmentation of cone beam computed tomography (CBCT) images. Although this method is widely used in clinical practice, it has some obvious limitations:

[0003] 1. Inaccuracy of manual segmentation: Manual segmentation relies on the operator’s experience and subjective judgment, which is highly subjective and may lead to poor repeatability and large errors in segmentation results.

[0004] 2. Low efficiency: The manual segmentation process is time-consuming, requiring an average of 1,390 seconds, which cannot meet the needs of rapid clinical evaluation. It not only increases the workload, but also prolongs the time patients have to wait for results.

[0005] 3. High cost: Manual segmentation requires a lot of manpower and time resources, which is costly. Summary of the invention

[0006] The purpose of this application is to provide a postoperative bone transplant material evaluation method, device, equipment, medium and product to solve the problems of low accuracy, long time consumption and high cost of traditional evaluation methods.

[0007] To achieve the above objectives, this application provides the following solutions:

[0008] In a first aspect, the present application provides a method for evaluating postoperative bone transplant materials, comprising:

[0009] Acquire historical CBCT images at different time points, and mark the historical CBCT images to determine the marked CBCT images; the marked CBCT images include the marked CBCT images on the day after surgery and the marked CBCT images after a set time period after surgery;

[0010] Based on the marked CBCT images at different time nodes, the historical CBCT images are used as input, and the marked bone graft material segmentation results are used as output, a neural network model is constructed and trained, and a trained neural network model is determined;

[0011] Inputting the CBCT images to be tested at different time nodes into the trained neural network, and outputting the bone graft material segmentation results;

[0012] According to the bone graft material segmentation results corresponding to different time nodes, the change degree of the bone graft material in the CBCT image to be tested is evaluated.

[0013] In a second aspect, the present application provides a postoperative bone transplant material evaluation device, comprising:

[0014] A marking module is used to obtain historical CBCT images at different time points, mark the historical CBCT images, and determine the marked CBCT images; the marked CBCT images include the marked CBCT images on the day after surgery and the marked CBCT images after a set time period after surgery;

[0015] A construction and training module is used to construct and train a neural network model based on the marked CBCT images at different time nodes, taking the historical CBCT images as input and the marked bone graft material segmentation results as output, and determine the trained neural network model;

[0016] A bone graft material segmentation result output module is used to input the CBCT images to be tested at different time nodes into the trained neural network and output the bone graft material segmentation result;

[0017] The change degree evaluation module is used to evaluate the change degree of the bone transplant material in the CBCT image to be tested according to the bone transplant material segmentation results corresponding to different time nodes.

[0018] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-described postoperative bone graft material evaluation methods.

[0019] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described postoperative bone transplant material evaluation methods.

[0020] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned postoperative bone transplant material evaluation methods.

[0021] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0022] The present application marks historical cone beam computed tomography (CBCT) images at different time nodes, and trains a neural network model based on the marked CBCT images to obtain the relationship between the CBCT images at different time nodes and the marked bone graft material segmentation results. Based on the relationship between the CBCT images and the bone graft material segmentation results, the present application uses the trained neural network model to predict the bone graft material segmentation results corresponding to the CBCT image to be tested, and evaluates the degree of change of the bone graft material. There is no need to manually segment the bone graft material in the CBCT image to be tested, that is, the bone graft material in the CBCT image to be tested can be automatically segmented, which greatly reduces the segmentation error and labor cost and improves the segmentation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 A flow chart of a method for evaluating postoperative bone transplant materials provided in one embodiment of the present application;

[0025] Figure 2 A schematic diagram of a computer device provided in accordance with an embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0027] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0028] The present application embodiment provides a method for evaluating postoperative bone graft materials. The method is executed by a computer device, and specifically can be executed by a computer device such as a terminal or a server alone, or can be executed by a terminal and a server together. In the present application embodiment, Figure 1 As shown, the method includes the following steps.

[0029] S1: Acquire historical CBCT images at different time points, mark the historical CBCT images, and determine the marked CBCT images; the marked CBCT images include the marked CBCT images on the day after surgery and the marked CBCT images after a set time period after surgery.

[0030] S2: Based on the marked CBCT images at different time nodes, the historical CBCT images are used as input and the marked bone graft material segmentation results are used as output to construct and train a neural network model and determine the trained neural network model.

[0031] S3: Inputting the CBCT images to be tested at different time nodes into the trained neural network, and outputting the bone transplant material segmentation result.

[0032] S4: According to the bone graft material segmentation results corresponding to different time nodes, the change degree of the bone graft material in the CBCT image to be tested is evaluated.

[0033] In an exemplary embodiment, a computer system is used to collect and preprocess image data: cone beam computed tomography (CBCT) images after maxillary sinus lift and implant surgery are collected and preprocessed to meet model input requirements.

[0034] In the actual application, 150 CBCT images after maxillary sinus lift and implant surgery were obtained from the hospital database, including images taken immediately after surgery (T0) and images taken 6 months after surgery (T1). All images were saved in DICOM format.

[0035] In an exemplary embodiment, S2 may be replaced by the following steps.

[0036] S21: Divide the historical CBCT images into data sets to determine a training set and a test set.

[0037] S22: Preprocess the historical CBCT images in the training set to determine the preprocessed CBCT images.

[0038] S23: Input the preprocessed CBCT image into the neural network model, use the feature extraction network backbone to extract the deep features of the preprocessed CBCT image, and generate a feature map; the neural network model includes a feature extraction network backbone, a decoder head and a loss function; the feature extraction network backbone is Swin Transformer; the decoder head is a UPerNet structure.

[0039] S24: Utilizing the decoder head, upsampling and fusing feature maps at different levels in a top-down manner combined with lateral connections, to determine a bone transplant material segmentation map.

[0040] S25: Optimizing the bone transplant material segmentation result in the bone transplant material segmentation map using a loss function to determine a trained neural network model.

[0041] In practical applications, the collected CBCT images are trained using the Swin-UPerNet model to automatically segment bone graft materials.

[0042] CBCT images can be labeled automatically or manually by a trained implantologist, providing the location and boundaries of the bone graft material on each slice. These labeled images serve as reference standards or “broad truth” data when training the model.

[0043] In this embodiment, 120 CBCT scan data are used to train the Swin-UPerNet model.

[0044] During the model training process, the specific steps are as follows:

[0045] Preprocessing of input images: The input images are resized to 512x512 pixels but not cropped to ensure the consistency of the model input while retaining as many image details as possible.

[0046] Epochs: The total number of training cycles for the model is set to 50 epochs. This means that the entire training set will be traversed 50 times to ensure that the model has enough time to learn the patterns in the data.

[0047] Adjustment of learning rate: The initial learning rate is set to 0.001 (1e-3), and is reduced by half at the 20th and 40th epochs. Adjustment of learning rate helps the model converge quickly in the early stages of training, and in the later stages, the model parameters are refined by reducing the step size to obtain a more accurate model.

[0048] Performance and inference speed evaluation: The performance and inference speed of all models are evaluated on a Tesla V100 GPU with 32GB of memory. This ensures that the evaluation process is efficient and repeatable.

[0049] Swin-UPerNet model architecture: The architecture of the Swin-UPerNet model consists of four parts: the slice image input, the backbone of the feature extraction network, the decoder head, and the loss function of the model.

[0050] Input slice image: The model receives preprocessed 512x512 pixel CBCT image slices as input.

[0051] Feature extraction network backbone: Swin Transformer is used as the backbone network of the model, which is responsible for extracting deep features of the image.

[0052] Decoder head: The decoder part adopts the UPerNet structure, which combines lateral connections in a top-down manner to achieve upsampling and fusion of feature maps to generate high-resolution segmentation maps.

[0053] Loss function: The model uses a specific loss function (such as Lovasz Loss) to optimize the segmentation results, especially when dealing with problems with unbalanced categories or unclear boundaries.

[0054] In an exemplary embodiment, after S25, the following steps are further included:

[0055] S26: Using the test set, evaluating the performance indicators of the trained neural network model.

[0056] S27: Adjusting the trained neural network model according to the performance indicator.

[0057] In practical applications, based on the deep learning server, the performance of the trained model on an independent test set is verified to ensure the accuracy and generalization ability of the model.

[0058] Furthermore, the model was tested using the remaining 30 CBCT scan data to evaluate the model's performance indicators, including 95% Hausdorff distance (HD), intersection over Union (IoU), Dice Similarity Coefficient (Dice), precision, accuracy, and volume deviation rate (VDR).

[0059] Of these, 95% are HD:

[0060] formula:

[0061] Explanation: 95% HausdorffDistance is the 95th percentile of the maximum distance between the automatic segmentation result and the ground truth data. This indicator measures the maximum deviation between the segmentation result and the true boundary. Among them, A is the predicted segmentation area, B is the true segmentation area, a is the voxel in the predicted segmentation area, b is the voxel in the true segmentation area, and ||||2 is the two norm.

[0062] IoU:

[0063] formula:

[0064] Explanation: IoU measures the degree of overlap between the automatic segmentation result and the true annotation. Among them, TP is the true positive, TN is the true negative, that is, the number of voxels correctly predicted as non-bone graft material, FP is the false positive, and FN is the false negative, that is, the number of voxels incorrectly predicted as non-bone graft material.

[0065] Dice:

[0066] formula:

[0067] Explanation: The Dice coefficient is another indicator to measure the accuracy of segmentation. It evaluates the similarity of segmentation by calculating the probability that two randomly selected sample points belong to both the predicted area and the true area.

[0068] Precision:

[0069] formula:

[0070] Explanation: Precision measures the proportion of samples correctly identified as bone graft materials in the automatic segmentation results, that is, the proportion of samples predicted to be positive that are actually positive.

[0071] Accuracy:

[0072] formula:

[0073] Explanation: Accuracy measures the proportion of correctly identified samples among all samples predicted to be positive or negative in the automatic segmentation results.

[0074] VDR:

[0075] formula:

[0076] Explanation: VDR measures the degree of deviation between the volume of bone graft material predicted by automatic segmentation and the actual volume.

[0077] These indicators are used together to evaluate the performance of the automatic segmentation model to ensure that the model can accurately identify and segment bone graft materials. Through the evaluation of these comprehensive indicators, we can fully understand the performance of the model in different aspects.

[0078] In an exemplary embodiment, S4 may be replaced by the following steps.

[0079] S41: Processing the bone transplant material segmentation results corresponding to different time nodes to determine the processed segmentation results.

[0080] S42: Based on the processed segmentation result, the voxels of the bone transplant material in the segmented area are counted to determine the volume of the bone transplant material corresponding to different time nodes.

[0081] S43: Compare the bone graft material volumes corresponding to different time nodes to evaluate the degree of change of the bone graft material in the CBCT image to be tested.

[0082] In practical applications, based on a computer system, the trained Swin-UPerNet model is used to automatically segment new CBCT images, identify bone graft materials, and evaluate changes in bone graft materials. If the volume of bone graft materials is reduced to the point where it cannot maintain the stability of the implant, bone grafting surgery is performed again or the implant is removed based on the clinical situation.

[0083] Furthermore, the CBCT image to be tested is input into the Swin-UPerNet model, and the model will automatically segment the bone graft material image. During operation, the backbone network of the Swin-UPerNet model extracts the features of the image; the decoder head fuses the feature maps at different levels to enhance the accuracy of segmentation. The model outputs the classification result of each pixel, that is, predicts whether each voxel in the image belongs to the bone graft material. The segmentation results output by the model are post-processed, such as removing small isolated areas and smoothing boundaries, to improve the segmentation quality.

[0084] Specific steps to assess changes in bone graft material:

[0085] 1. Volume calculation: Using the automatic segmentation results, calculate the volume of the bone graft material. This can be achieved by counting the voxels within the segmented area.

[0086] 2. Comparative analysis: Compare the CBCT image segmentation results at different time points (such as immediately after surgery and 6 months after surgery) to evaluate the changes in the volume of bone graft material.

[0087] 3. Statistical analysis: Use statistical methods, such as paired t-test, to evaluate the statistical significance of changes in bone graft material volume.

[0088] 4. Visualization: Through 3D reconstruction technology, the segmentation results are visualized to intuitively display the position and volume changes of bone graft materials.

[0089] 6. Accuracy evaluation: Use indicators such as 95% Hausdorff distance, IoU, and Dice similarity coefficient to evaluate the accuracy of the automatic segmentation results.

[0090] 7. Clinical correlation analysis: Combined with clinical data, the correlation between changes in bone graft materials and clinical outcomes such as implant survival rate and patient recovery is analyzed.

[0091] 8. Report generation: Based on the above analysis, a detailed report is generated, which includes segmentation results, volume change assessment and clinical relevance analysis.

[0092] In practical applications, based on the computer system, the accuracy of the automatic segmentation results is evaluated, and the model is optimized as needed, that is, the accuracy of the segmentation results is evaluated using performance indicators such as 95% HD, IoU, Dice, etc., and the model parameters are adjusted according to the evaluation results to optimize the performance.

[0093] In an exemplary embodiment, S41 may be replaced by the following steps.

[0094] S411: removing isolated areas in the bone transplant material segmentation results corresponding to different time nodes, and smoothing the boundaries of the bone transplant material segmentation results corresponding to different time nodes to determine the processed segmentation results.

[0095] In an exemplary embodiment, after S4, the following steps are further included:

[0096] S5: Determine whether to re-perform bone grafting surgery or remove the implant according to the degree of change of the bone graft material.

[0097] In practical applications, based on a computer system, the processing time of the automatic segmentation method is compared with that of the manual segmentation method to verify the efficiency advantage of the automatic method.

[0098] In the present application, the same CBCT image to be tested is processed based on the above technical solution. After testing, the processing time of the automatic segmentation method (19.28 seconds) is recorded and compared with the average processing time of the manual segmentation method (1390 seconds). It can be seen that the processing time of the present application is much shorter than the processing time of the manual segmentation method, and the segmentation efficiency of the present application is significantly higher than the segmentation efficiency of the manual segmentation method.

[0099] Based on the same inventive concept, the embodiment of the present application also provides a postoperative bone transplant material evaluation device for implementing the postoperative bone transplant material evaluation method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more postoperative bone transplant material evaluation device embodiments provided below can refer to the limitations of the postoperative bone transplant material evaluation method above, and will not be repeated here.

[0100] In an exemplary embodiment, a postoperative bone graft material evaluation device is provided, comprising:

[0101] The marking module is used to obtain historical CBCT images at different time nodes, mark the historical CBCT images, and determine the marked CBCT images; the marked CBCT images include the marked CBCT images on the day after surgery and the marked CBCT images after a set time period after surgery.

[0102] A construction and training module is used to construct and train a neural network model based on the marked CBCT images at different time nodes, taking the historical CBCT images as input and the marked bone graft material segmentation results as output, and determine the trained neural network model.

[0103] The bone transplant material segmentation result output module is used to input the CBCT images to be tested at different time nodes into the trained neural network and output the bone transplant material segmentation result.

[0104] The change degree evaluation module is used to evaluate the change degree of the bone transplant material in the CBCT image to be tested according to the bone transplant material segmentation results corresponding to different time nodes.

[0105] Currently, there is no artificial intelligence-based automatic segmentation method for bone graft materials after maxillary sinus lift and simultaneous implantation on the market, that is, the postoperative bone graft material evaluation method or device mentioned above in this application. This application can simultaneously achieve high-precision and high-efficiency bone graft material segmentation.

[0106] In practical applications, this application determines the bone transplant material segmentation results based on the Swin-UPerNet model, which can improve accuracy: through the high-precision segmentation capability of the deep learning model, human errors are reduced and the accuracy of the results is improved.

[0107] Improve efficiency: The automatic segmentation method reduces the time for automatic segmentation of a single CBCT image from 1390 seconds to 19.28 seconds, significantly improving the evaluation efficiency.

[0108] Reduce costs: Reduce reliance on manpower, reduce operating costs, and make clinical evaluation more cost-effective.

[0109] In an exemplary embodiment, a computer device is provided, such as Figure 2As shown, the computer device can be a server or a terminal. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store postoperative bone transplant material evaluation data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a postoperative bone transplant material evaluation method is implemented.

[0110] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the above method is implemented when the processor executes the computer program.

[0111] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the above method when executed by a processor.

[0112] In an exemplary embodiment, a computer program product is provided, including a computer program, which implements the above method when executed by a processor.

[0113] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ReadOnlyMemory, ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (Magnetoresistive RandomAccess Memory, MRAM), ferroelectric random access memory (Ferroelectric RandomAccess Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (RandomAccess Memory, RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0114] In this application, all actions to obtain signals, information or data are carried out in compliance with the relevant data protection laws and policies of the country where they are located and with the authorization given by the owner of the corresponding device.

[0115] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0116] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0117] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for evaluating postoperative bone transplant materials, characterized in that: The postoperative bone graft material evaluation method comprises: Acquire historical CBCT images at different time points, and mark the historical CBCT images to determine the marked CBCT images; the marked CBCT images include the marked CBCT images on the day after surgery and the marked CBCT images after a set time period after surgery; Based on the marked CBCT images at different time nodes, the historical CBCT images are used as input, and the marked bone graft material segmentation results are used as output, a neural network model is constructed and trained, and a trained neural network model is determined; Inputting the CBCT images to be tested at different time nodes into the trained neural network, and outputting the bone graft material segmentation results; According to the bone graft material segmentation results corresponding to different time nodes, the change degree of the bone graft material in the CBCT image to be tested is evaluated.

2. The postoperative bone graft material evaluation method according to claim 1, characterized in that: Based on the marked CBCT images at different time nodes, the historical CBCT images are used as input and the marked bone graft material segmentation results are used as output, a neural network model is constructed and trained, and the trained neural network model is determined, specifically including: Dividing the historical CBCT images into data sets to determine a training set and a test set; Preprocessing the historical CBCT images in the training set to determine the preprocessed CBCT images; The preprocessed CBCT image is input into the neural network model, and the deep features of the preprocessed CBCT image are extracted by using the feature extraction network backbone to generate a feature map; the neural network model includes a feature extraction network backbone, a decoder head and a loss function; the feature extraction network backbone is a Swin Transformer; and the decoder head is a UPerNet structure; Using the decoder head, upsampling and fusing feature maps at different levels in a top-down manner combined with lateral connections to determine a bone graft material segmentation map; The loss function is used to optimize the bone graft material segmentation result in the bone graft material segmentation map to determine the trained neural network model.

3. The postoperative bone graft material evaluation method according to claim 2, characterized in that: The loss function is used to optimize the bone graft material segmentation result in the bone graft material segmentation map to determine the trained neural network model, and then the following steps are further included: Using the test set, evaluating the performance indicators of the trained neural network model; The trained neural network model is adjusted according to the performance indicator.

4. The postoperative bone graft material evaluation method according to claim 1, characterized in that: According to the bone graft material segmentation results corresponding to different time nodes, evaluating the degree of change of the bone graft material in the CBCT image to be tested, specifically including: Processing the bone transplant material segmentation results corresponding to different time nodes to determine the processed segmentation results; Based on the processed segmentation results, the voxels of the bone graft material in the segmented area are counted to determine the volume of the bone graft material corresponding to different time nodes; The bone graft material volumes corresponding to different time points are compared to evaluate the degree of change of the bone graft material in the CBCT image to be tested.

5. The postoperative bone graft material evaluation method according to claim 4, characterized in that: The bone graft material segmentation results corresponding to different time nodes are processed to determine the processed segmentation results, specifically including: The isolated areas in the bone graft material segmentation results corresponding to different time nodes are removed, and the boundaries of the bone graft material segmentation results corresponding to different time nodes are smoothed to determine the processed segmentation results.

6. The postoperative bone graft material evaluation method according to claim 1, characterized in that: According to the bone graft material segmentation results corresponding to different time nodes, the change degree of the bone graft material in the CBCT image to be tested is evaluated, and then the following steps are further included: Whether to perform bone grafting again or remove the implant is determined based on the degree of change of the bone graft material.

7. A postoperative bone transplant material evaluation device, characterized in that: The postoperative bone transplant material evaluation device comprises: A marking module is used to obtain historical CBCT images at different time points, mark the historical CBCT images, and determine the marked CBCT images; the marked CBCT images include the marked CBCT images on the day after surgery and the marked CBCT images after a set time period after surgery; A construction and training module is used to construct and train a neural network model based on the marked CBCT images at different time nodes, taking the historical CBCT images as input and the marked bone graft material segmentation results as output, and determine the trained neural network model; A bone graft material segmentation result output module is used to input the CBCT images to be tested at different time nodes into the trained neural network and output the bone graft material segmentation result; The change degree evaluation module is used to evaluate the change degree of the bone transplant material in the CBCT image to be tested according to the bone transplant material segmentation results corresponding to different time nodes.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the postoperative bone graft material evaluation method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the postoperative bone graft material evaluation method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the postoperative bone graft material evaluation method according to any one of claims 1 to 6 is implemented.

Images