Semantic medical images for 3D printing anatomical structures

Abstract

The invention relates to the technical field of medical image processing, in particular to a semantic medical image for 3D printing of an anatomical structure, which comprises the following steps: acquiring medical image data, performing automatic labeling and classification on the anatomical structure in the medical image data through a semantic segmentation model to obtain anatomical structure labeling information, according to the obtained anatomical structure labeling information, an image reconstruction algorithm is adopted to conduct three-dimensional reconstruction on each anatomical structure to obtain a three-dimensional mesh model, the three-dimensional mesh model is optimized on the basis of the obtained three-dimensional mesh model in combination with material characteristics and a mechanical model to obtain optimized three-dimensional printing model data, and a 3D printing technology is adopted to obtain the 3D printing model. And the three-dimensional printing model data are input into printing equipment, 3D printing of the anatomical structure is completed, and a physical model of the anatomical structure is obtained. According to the invention, the accuracy of medical image analysis is improved, and the precision and printing quality of the three-dimensional model are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to using semantic medical images for 3D printing of anatomical structures. Background Art

[0002] With the rapid development of medical imaging technology, especially the widespread application of CT and MRI technology, the acquisition of medical imaging data has become more accurate. Medical imaging provides detailed two-dimensional slice information. Two-dimensional images can only display the plane information inside the human body, and it is difficult to intuitively present the complex three-dimensional anatomical structure of the human body. This limitation makes doctors often face the problem of insufficient anatomical structure information when making diagnoses and surgical planning. Therefore, three-dimensional reconstruction technology was proposed and widely used in the medical field.

[0003] However, existing 3D reconstruction technology still faces a series of technical challenges. The automatic labeling and classification of anatomical structures in medical images is a technical problem. Although the application of deep learning in image processing has made significant progress, how to effectively and accurately automatically label complex anatomical structures (such as bones, organs and blood vessels) in medical images is still an urgent problem to be solved. Although some reconstruction algorithms can achieve preliminary 3D reconstruction, due to the complexity of anatomical structures, existing 3D reconstruction technology often cannot effectively solve data distortion and errors, resulting in the reconstructed 3D model being inaccurate or having certain deformations. Especially in the 3D printing process, these precision problems may affect the usability of the final printed anatomical model. Summary of the invention

[0004] The present invention provides semantic medical images for 3D printing of anatomical structures.

[0005] Semantic medical images are used to 3D print anatomical structures, including the following steps:

[0006] S1: acquiring medical image data, the medical image data including CT scan image data and MRI scan image data, and preprocessing the medical image data;

[0007] S2: Automatically labeling and classifying anatomical structures in medical image data through a semantic segmentation model to obtain anatomical structure labeling information, wherein the anatomical structures include bones, organs, and blood vessels;

[0008] S3: Based on the obtained anatomical structure annotation information, an image reconstruction algorithm is used to perform three-dimensional reconstruction on each anatomical structure to obtain a three-dimensional mesh model;

[0009] S4: Based on the obtained three-dimensional mesh model, the three-dimensional mesh model is optimized in combination with material properties and mechanical models to obtain optimized three-dimensional printing model data;

[0010] S5: Using 3D printing technology, the three-dimensional printing model data is input into the printing device to complete the 3D printing of the anatomical structure and obtain the physical model of the anatomical structure.

[0011] Optionally, the S1 includes:

[0012] S11: Acquire medical image data through a medical device, wherein the medical image data includes CT scan data and MRI scan data, and the medical device includes a CT scanner and an MRI scanner;

[0013] S12: performing spatial standardization processing on the acquired medical imaging data;

[0014] S13: performing noise removal processing on the acquired medical image data;

[0015] S14: performing contrast enhancement processing on the denoised medical image data.

[0016] Optionally, the S2 specifically includes:

[0017] S21: Select DeepLabV3+ as the semantic segmentation model;

[0018] S22: Use annotated medical imaging data (such as CT or MRI scans) to train a semantic segmentation model;

[0019] S23, after the model training is completed, the semantic segmentation model infers the new medical image data and automatically annotates the different anatomical structures in the medical image data;

[0020] S24, after the trained semantic segmentation model is annotated and classified in the medical image data, the annotation information of each anatomical structure is output.

[0021] Optionally, the S3 includes:

[0022] S31: performing three-dimensional reconstruction of each anatomical structure using a voxel reconstruction algorithm according to the anatomical structure annotation information obtained in step S2;

[0023] S32: In the voxel reconstruction algorithm, a multi-level reconstruction strategy is adopted to perform three-dimensional reconstruction in layers according to the complexity of different anatomical structures;

[0024] S33: Generate three-dimensional voxel data through voxel reconstruction algorithm and multi-level reconstruction strategy, and convert the three-dimensional voxel data into a three-dimensional mesh model.

[0025] Optionally, S3 further includes:

[0026] S34, by comparing with the three-dimensional anatomical data in a standard anatomical database (such as the Visible Human Project), the distorted area in the three-dimensional mesh model is identified and corrected.

[0027] Optionally, the S4 includes:

[0028] S41, based on 3D printing requirements, select the type of printing materials, including biocompatible materials and polymers, ensure that the selected materials meet the requirements of strength, elasticity and biocompatibility of the printed anatomical structures, and evaluate the material properties according to the selected material type;

[0029] S42, based on the material properties, a finite element analysis model is used to perform mechanical analysis on the three-dimensional mesh model.

[0030] S43: According to the material properties and mechanical analysis results, the 3D mesh model is optimized to optimize the structure, thickness and geometry of the 3D mesh model.

[0031] Optionally, S4 further includes:

[0032] S44, performing printing feasibility evaluation on the optimized three-dimensional mesh model;

[0033] S45, optimize the thin-walled parts in the 3D mesh model while ensuring the accuracy of the anatomical structure to avoid fractures caused by mechanical instability during printing;

[0034] S46, converting the optimized three-dimensional mesh model into a file format for 3D printing (such as STL, OBJ, etc.) to obtain three-dimensional printing model data.

[0035] Optionally, the S5 includes:

[0036] S51, selecting a suitable 3D printing technology for printing according to the optimized 3D printing model data;

[0037] S52, according to the selected 3D printing technology, setting the parameters of the printer, such as printing layer thickness, printing speed, printing temperature, etc., to ensure the stability and accuracy of the printing process;

[0038] S53, inputting the three-dimensional printing model data into the 3D printing device, and printing according to the set printing parameters.

[0039] Optionally, S5 further includes:

[0040] S54, after printing is completed, post-processing the 3D printed model of the anatomical structure;

[0041] S55 performs quality inspection on the completed 3D printed models to ensure that they meet the expected standards in terms of shape, size, details and accuracy.

[0042] Beneficial effects of the present invention:

[0043] The present invention, by combining deep learning technology, uses a semantic segmentation model (such as DeepLabV3+) to automatically label and classify anatomical structures (such as bones, organs, and blood vessels) in medical image data, significantly improving the labeling accuracy. Through this process, the spatial position and morphology of each anatomical structure can be accurately captured. This accuracy improvement not only ensures that the details in the medical image data can be fully preserved, but also provides high-quality basic data for subsequent three-dimensional reconstruction, printing, and medical applications, greatly enhancing the practicality in clinical surgery, medical education, and pathological analysis.

[0044] The present invention optimizes the three-dimensional mesh model by combining material properties and mechanical models, thereby ensuring the stability and reliability of the printed model in terms of mechanical properties. By introducing finite element analysis to evaluate mechanical properties and evaluate printing feasibility, the present invention achieves accurate material selection and optimization of mechanical properties. The optimized three-dimensional model is more in line with actual application requirements during the printing process, ensuring that the printed anatomical model not only meets medical standards in terms of accuracy, but can also meet the requirements of medical surgery simulation, education and training, clinical treatment, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 The figure is a schematic diagram of a method flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0048] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0049] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0050] like Figure 1 As shown, semantic medical images are used to 3D print anatomical structures, including the following steps:

[0051] S1: acquiring medical image data, the medical image data including CT scan image data and MRI scan image data, and preprocessing the medical image data;

[0052] S2: The anatomical structures in the medical image data are automatically labeled and classified through the semantic segmentation model to obtain the anatomical structure annotation information. The anatomical structures include bones, organs and blood vessels.

[0053] S3: Based on the obtained anatomical structure annotation information, an image reconstruction algorithm is used to perform three-dimensional reconstruction on each anatomical structure to obtain a three-dimensional mesh model;

[0054] S4: Based on the obtained three-dimensional mesh model, the three-dimensional mesh model is optimized in combination with material properties and mechanical models to obtain optimized three-dimensional printing model data;

[0055] S5: Using 3D printing technology, the three-dimensional printing model data is input into the printing device to complete the 3D printing of the anatomical structure and obtain the physical model of the anatomical structure.

[0056] S1 includes:

[0057] S11: acquiring medical imaging data through medical equipment, where the medical imaging data includes CT scanning data and MRI scanning data, and the medical equipment includes a CT scanner and an MRI scanner;

[0058] S12: performing spatial standardization processing on the acquired medical imaging data;

[0059] In view of the different spatial resolutions and coordinate systems of CT scan data and MRI scan data, the nearest neighbor interpolation method is used to uniformly convert medical imaging data into the same resolution and coordinate system, thereby ensuring the uniformity and comparability of different data sources. The nearest neighbor interpolation method is expressed as:

[0060] I new (x,y)=Iold(x round ,y round );

[0061] Among them, I new (x, y) is the interpolation result at the new coordinate (x, y), I old (x round ,y round ) is the nearest neighbor point (x) in the original image. round ,y round ), (x round ,y round ) is the integer coordinate closest to the new coordinate point (x, y);

[0062] S13: Perform noise removal on the acquired medical image data. Use wavelet transform to remove high-frequency noise in the medical image data and retain important anatomical features and details in the medical image data to improve the accuracy of subsequent image analysis. The wavelet transform is expressed as:

[0063]

[0064] in, is the signal after wavelet transformation, f(t) is the original signal (i.e., medical imaging data), ψ(ta) is the wavelet basis function, a is the translation factor, and t is the time or space variable;

[0065] S14: performing contrast enhancement processing on the denoised medical image data, using histogram equalization to enhance the details in the medical image data, especially at the boundaries of anatomical structures, to enhance the visibility and contrast of key areas;

[0066]

[0067] Among them, s k is the equalization result of the pixel value L in the image, L is the maximum pixel value (for example, L = 256 in an 8-bit image), N is the total number of pixels in the image, and p i is the frequency of occurrence of pixel value i in the image, is the cumulative probability that the pixel value is less than or equal to k.

[0068] S2 specifically includes:

[0069] S21: DeepLabV3+ is selected as the semantic segmentation model. DeepLabV3+ is a deep learning architecture based on convolutional neural networks. It is widely used in semantic segmentation tasks. It uses technologies such as dilated convolution and spatial pyramid pooling. It can efficiently capture contextual information of different scales in images and is suitable for anatomical structure segmentation tasks in medical images.

[0070] S22: training a semantic segmentation model using annotated medical imaging data (such as CT or MRI scan images), where the annotated medical imaging data includes annotated anatomical structures (such as bones, organs, and blood vessels);

[0071] The cross entropy loss function is used as the loss function during training, and its formula is:

[0072]

[0073] Where L is the number of categories (i.e., bones, organs, blood vessels), L is the true label of the pixel belonging to category i (usually 0 or 1), and p i is the probability that the model predicts that the pixel belongs to category i;

[0074] This loss function aims to minimize the difference between the true label and the model prediction, thereby improving the segmentation accuracy;

[0075] S23, after the model training is completed, the semantic segmentation model infers the new medical image data and automatically labels the different anatomical structures (e.g., bones, organs, or blood vessels) in the medical image data. The model assigns a category label to each pixel and generates a segmented region for each anatomical structure.

[0076] Classification tasks: The model categorizes each pixel into different anatomical structure categories, such as bones, blood vessels, or organs;

[0077] Labeling tasks: By assigning pixel-level category labels, the model can automatically label the specific location and morphology of anatomical structures, thereby providing accurate data for subsequent 3D reconstruction and printing;

[0078] S24, after the trained semantic segmentation model is annotated and classified in the medical image data, the annotation information of each anatomical structure is output, including:

[0079] Class labels for each pixel: for example, bone, heart, and blood vessels;

[0080] Format of annotation information: Output is a structured data format, such as a two-dimensional label image (each pixel corresponds to a category label), or outputs the location, size, and other information of each structure in other formats (such as JSON, XML);

[0081] These annotation information will provide necessary input for subsequent steps such as 3D reconstruction, physical model optimization and 3D printing.

[0082] S3 includes:

[0083] S31: Based on the anatomical structure annotation information obtained in step S2, a voxel reconstruction algorithm is used to perform three-dimensional reconstruction of each anatomical structure. The voxel reconstruction algorithm converts two-dimensional medical image data into three-dimensional volume data to achieve multi-level and refined reconstruction of the anatomical structure;

[0084] Voxelization process: First, the two-dimensional annotation information (such as segmentation boundaries) of each anatomical structure (such as bones, organs, blood vessels, etc.) is converted into a three-dimensional voxel grid. In this process, each pixel in the annotation information is mapped to the three-dimensional grid according to its spatial coordinate position and assigned corresponding attribute values;

[0085] Interpolation method: In order to ensure the accuracy of the anatomical structure during the reconstruction process, the cubic convolution interpolation method is used to spatially interpolate the image data, seamlessly connecting the anatomical structures between different slices to obtain a high-resolution three-dimensional structure;

[0086] S32: In the voxel reconstruction algorithm, a multi-level reconstruction strategy is adopted to perform three-dimensional reconstruction in layers according to the complexity of different anatomical structures. This strategy can improve the accuracy of reconstruction details and ensure that the spatial position, morphological characteristics and mutual relationships of different structures are accurately captured.

[0087] Hierarchical reconstruction: Different reconstruction resolutions are set for each anatomical structure based on its size, shape, and complexity. Simpler structures (such as bones) can be reconstructed at a lower resolution, while structures with complex details (such as blood vessels) are reconstructed at a higher resolution to ensure accuracy;

[0088] Reconstruction accuracy control: By automatically adjusting the voxel size of each layer during the reconstruction process, it ensures that reasonable details can be obtained in different parts of the anatomical structure and avoids overfitting or data sparsity;

[0089] S33: Generate 3D voxel data through voxel reconstruction algorithm and multi-level reconstruction strategy, and convert the 3D voxel data into a 3D mesh model;

[0090] The generated 3D voxel data is converted into a 3D mesh model (such as STL format) through a surface reconstruction algorithm (such as Marching Cubes). The mesh model consists of a set of triangular facets that can accurately represent the shape of the anatomical structure and is suitable for subsequent 3D printing, surgical planning or other medical applications.

[0091] Mesh generation process: First, extract the surface of the 3D voxel data and convert the voxel data into a continuous surface mesh using the Marching Cubes algorithm. Then, optimize the generated 3D mesh model to ensure its smoothness, physical authenticity and printing feasibility. Finally, a fine mesh model that can be used for 3D printing is obtained.

[0092] S3 also includes:

[0093] In the process of 3D reconstruction of anatomical structures, the resolution limitation of the image acquisition device, the change of scanning angle and the interpolation error in the reconstruction algorithm may cause the final generated 3D mesh model to have shape distortion, errors and inaccurate parts. Therefore, it is necessary to effectively correct these distortions and errors according to the standards in the medical field to ensure that the final 3D mesh model is highly consistent with the actual anatomical structure in terms of the representation of the anatomical structure.

[0094] S34, by comparing with the 3D anatomical data in a standard anatomical database (such as the Visible Human Project), the distorted area in the 3D mesh model is identified and corrected. This process relies on the high-precision 3D standard model in the database as a comparison benchmark to correct the generated 3D mesh model. The comparison steps are as follows:

[0095] The generated 3D mesh model is spatially aligned with the reference model in the standard anatomical database to ensure that the coordinate systems of the two are consistent;

[0096] Use the least squares method to compare the shape of the mesh model, optimize its geometry, reduce errors, and correct distorted areas;

[0097] The corresponding point set between the generated 3D mesh model and the reference model in the standard anatomical database is set as {P i} and {Q i}, where i = 1, 2, ..., N;

[0098] The goal of the least squares method is to minimize the difference between two sets of points, expressed as:

[0099]

[0100] Through the optimization algorithm, the geometric position of the 3D mesh model is adjusted to minimize the error E, thereby correcting the distortion part;

[0101] After correction, iterative optimization is used to ensure that the generated three-dimensional mesh model is highly consistent with the anatomical structure in the standard database and meets the standard requirements in the medical field.

[0102] S4 includes:

[0103] S41, based on 3D printing requirements, select the type of printing material, including biocompatible materials and polymers, to ensure that the selected materials meet the strength, elasticity and biocompatibility requirements of the printed anatomical structure, and evaluate the material properties according to the selected material type, especially the analysis of elastic modulus and tensile strength, to ensure that the printed 3D model can accurately simulate the behavior of human tissue in mechanics and biology.

[0104] Elastic modulus: used to describe the ease with which a material deforms. The formula is:

[0105]

[0106] Among them, E is the elastic modulus, σ is the stress, and ε is the strain;

[0107] Tensile strength: used to measure the maximum stress a material can withstand during stretching until it breaks;

[0108] S42, based on the material properties, uses the finite element analysis model to perform mechanical analysis on the three-dimensional mesh model, including:

[0109] Finite element analysis is used to test the stress of the 3D mesh model and calculate the stress, strain and deformation of each part of the 3D mesh model to ensure that the printed anatomical structure can meet the actual application requirements.

[0110] By optimizing the calculation, the mechanical properties of the three-dimensional mesh model are improved to ensure that the model has sufficient mechanical strength and appropriate elasticity, avoiding problems of excessive rigidity or excessive fragility.

[0111] S43: According to the results of material properties and mechanical analysis, the three-dimensional mesh model is optimized to optimize the structure, thickness and geometric shape of the three-dimensional mesh model. The three-dimensional printing model data obtained after optimization will have geometric properties that meet the requirements of strength, elasticity, etc., and is suitable for 3D printing.

[0112] S4 also includes:

[0113] S44, conducts a printing feasibility assessment on the optimized 3D mesh model to ensure that it meets the printing accuracy and material requirements of the 3D printer, considers possible material shrinkage and deformation during the printing process, and ensures that no failure or unqualified results will occur during the printing process. During the assessment process, print path planning and support structure design are used, and the model's size, structural stability and other aspects are optimized.

[0114] Support structure design: According to the shape of the 3D model and the properties of the printing material, design an appropriate support structure to prevent material deformation or model breakage during printing. The design of the support structure needs to be reasonably laid out according to the printing range and angle of the printer to ensure that every part of the model can be effectively supported;

[0115] S45, optimize the thin-walled parts in the 3D mesh model while ensuring the accuracy of the anatomical structure to avoid fractures caused by mechanical instability during printing;

[0116] S46, converting the optimized three-dimensional mesh model into a file format for 3D printing (such as STL, OBJ, etc.) to obtain three-dimensional printing model data.

[0117] S5 includes:

[0118] S51, selecting a suitable 3D printing technology for printing according to the optimized 3D printing model data;

[0119] S52, according to the selected 3D printing technology, setting the parameters of the printer, such as printing layer thickness, printing speed, printing temperature, etc., to ensure the stability and accuracy of the printing process;

[0120] S53, inputting the 3D printing model data into the 3D printing device, printing according to the set printing parameters, and generating a physical model of the anatomical structure layer by layer. During the printing process, the device accurately deposits the material according to the printing path to ensure that the shape and details of each anatomical structure part are accurately reproduced.

[0121] The S5 also includes:

[0122] S54, after printing is completed, the 3D printed model of the anatomical structure is post-processed to improve the surface quality and mechanical properties, and the post-processing includes removing support structures, surface polishing and heat treatment;

[0123] S55 performs quality inspection on the completed 3D printed models to ensure that they meet the expected standards in terms of shape, size, details and accuracy.

[0124] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0125] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. Semantic medical images are used to 3D print anatomical structures, characterized in that: The following steps are involved: S1: acquiring medical image data, the medical image data including CT scan image data and MRI scan image data, and preprocessing the medical image data; S2: Automatically labeling and classifying anatomical structures in medical image data through a semantic segmentation model to obtain anatomical structure labeling information, wherein the anatomical structures include bones, organs, and blood vessels; S3: Based on the obtained anatomical structure annotation information, an image reconstruction algorithm is used to perform three-dimensional reconstruction on each anatomical structure to obtain a three-dimensional mesh model; S4: Based on the obtained three-dimensional mesh model, the three-dimensional mesh model is optimized in combination with material properties and mechanical models to obtain optimized three-dimensional printing model data; S5: Using 3D printing technology, the three-dimensional printing model data is input into the printing device to complete the 3D printing of the anatomical structure and obtain the physical model of the anatomical structure.

2. The semantic medical image according to claim 1 is used for 3D printing of anatomical structures, characterized in that: The S1 includes: S11: Acquire medical image data through a medical device, wherein the medical image data includes CT scan data and MRI scan data, and the medical device includes a CT scanner and an MRI scanner; S12: performing spatial standardization processing on the acquired medical imaging data; S13: performing noise removal processing on the acquired medical image data; S14: performing contrast enhancement processing on the denoised medical image data.

3. The semantic medical image for 3D printing of anatomical structures according to claim 2, characterized in that: The S2 specifically includes: S21: Select DeepLabV3+ as the semantic segmentation model; S22: Use labeled medical imaging data to train semantic segmentation models; S23, after the model training is completed, the semantic segmentation model infers the new medical image data and automatically annotates the different anatomical structures in the medical image data; S24, after the trained semantic segmentation model is annotated and classified in the medical image data, the annotation information of each anatomical structure is output.

4. The semantic medical image for 3D printing of anatomical structures according to claim 3, characterized in that: The S3 includes: S31: performing three-dimensional reconstruction of each anatomical structure using a voxel reconstruction algorithm according to the anatomical structure annotation information obtained in step S2; S32: In the voxel reconstruction algorithm, a multi-level reconstruction strategy is adopted to perform three-dimensional reconstruction in layers according to the complexity of different anatomical structures; S33: Generate three-dimensional voxel data through voxel reconstruction algorithm and multi-level reconstruction strategy, and convert the three-dimensional voxel data into a three-dimensional mesh model.

5. The semantic medical image for 3D printing of anatomical structures according to claim 4, characterized in that: The S3 further includes: S34, identifying the distorted area in the three-dimensional mesh model by comparing it with the three-dimensional anatomical data in the standard anatomical database, and performing correction.

6. The semantic medical image for 3D printing of anatomical structures according to claim 5, characterized in that: The S4 includes: S41, based on 3D printing requirements, select the type of printing materials, including biocompatible materials and polymers, and evaluate the material properties according to the selected material type; S42, based on the material properties, a finite element analysis model is used to perform mechanical analysis on the three-dimensional mesh model; S43: Optimize the 3D mesh model based on the material properties and mechanical analysis results.

7. The semantic medical image for 3D printing of anatomical structures according to claim 6, characterized in that: The S4 further comprises: S44, performing printing feasibility evaluation on the optimized three-dimensional mesh model; S45, optimizes the thin-walled parts in the 3D mesh model to avoid fractures caused by mechanical instability during printing; S46, converting the optimized three-dimensional mesh model into a file format for 3D printing to obtain three-dimensional printing model data.

8. The semantic medical image for 3D printing of anatomical structures according to claim 7, characterized in that: The S5 includes: S51, selecting 3D printing technology for printing according to the optimized 3D printing model data; S52, setting parameters of the printer according to the selected 3D printing technology; S53, inputting the three-dimensional printing model data into the 3D printing device, and printing according to the set printing parameters.

9. The semantic medical image for 3D printing of anatomical structures according to claim 8, characterized in that: The S5 further includes: S54, after printing is completed, post-processing the 3D printed model of the anatomical structure; S55, performing quality inspection on the 3D printed model that has been printed.

Images