A computer-aided based cranial prosthesis design and manufacturing system

Abstract

The present application relates to the technical field of computer-aided skull repair prosthesis design and manufacturing, in particular to a computer-aided skull repair prosthesis design and manufacturing system. The system comprises a shape prediction unit for solving optimal shape coefficients and generating an initial prosthesis shape; a constraint evaluation unit for calculating a prosthesis conflict index and dividing a conflict level; a correction field generation unit for responding to the prosthesis conflict index, decomposing a physical violation amount, and solving a dense displacement field covering all vertices of the prosthesis through a shape correction propagation model; and a closed-loop optimization unit for applying the dense displacement field to geometric deformation of the initial prosthesis shape to obtain a corrected prosthesis shape, and re-submitting the corrected prosthesis shape to the constraint evaluation unit to execute a hierarchical final version strategy. The present application solves the contradiction between anatomical fidelity and clinical safety constraints in skull repair by constructing a system comprising shape prediction, constraint evaluation, correction generation and closed-loop optimization.

Description

Technical Field

[0001] This invention relates to the field of computer-aided cranioplasty prosthesis design and manufacturing technology, specifically a computer-aided cranioplasty prosthesis design and manufacturing system. Background Technology

[0002] In the field of computer-aided design and manufacturing of cranioplasty prostheses, it is crucial to generate personalized implants for patients that combine high fidelity and clinical safety. Existing design methods, such as the traditional mirror method, mainly focus on restoring the natural anatomical shape and appearance symmetry of the skull to achieve ideal aesthetic results. However, this often results in the initial prosthesis failing to meet the rigid safety requirements of clinical surgery in key physical properties. For example, problems may arise such as excessively thin local thickness due to overfitting, insufficient safe distance from important nerves or vascular sinuses, or mismatch in edge curvature. This creates an inherent and irreconcilable conflict between the anatomical aesthetics of the prosthesis and the safety of clinical surgery.

[0003] In existing technologies, the process of resolving such conflicts often relies on designers' manual and repeated adjustments, which is not only inefficient, but also the quality of the final solution is highly dependent on personal experience and lacks an objective, quantitative evaluation and correction loop. Therefore, how to provide a cranioplasty prosthesis design method that can automatically resolve the conflict between anatomical high fidelity and multi-dimensional clinical safety constraints is a problem that urgently needs to be solved by those skilled in the art.

[0004] The information disclosed in the background section above is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention discloses a computer-aided cranial prosthesis design and manufacturing system. Specifically, the technical solution of this invention includes:

[0006] The morphological prediction unit is used to solve for the optimal morphological coefficients and generate the initial prosthesis morphology based on a preset statistical shape model and the collected patient skull reference surface.

[0007] The constraint assessment unit is used to perform geometric analysis on the initial prosthesis shape based on preset multi-dimensional surgical constraint limits, calculate the prosthesis conflict index, and classify the conflict level according to the prosthesis conflict index.

[0008] The modified field generation unit is used to deconstruct the physical violation quantity in response to the prosthesis conflict index, and solve for the generation of a dense displacement field covering all vertices of the prosthesis through the morphological modification propagation model.

[0009] The closed-loop optimization unit is used to apply a dense displacement field to geometrically deform the initial prosthesis shape to obtain the corrected prosthesis shape, and then resubmits the corrected prosthesis shape to the constraint evaluation unit to execute the hierarchical final version strategy.

[0010] Preferably, the statistical shape model includes an average skull shape vector and principal variation pattern constructed by morphological alignment and principal component analysis of a large-scale standard skull dataset; the patient skull reference surface is obtained by acquiring CT data of the patient's skull defect and automatically extracting the defect boundary and the remaining effective skull surface.

[0011] Preferably, the multi-dimensional surgical constraint limits include: minimum permissible strength thickness of the prosthesis, maximum appearance restriction thickness, minimum safe distance from critical nerves or vascular sinuses, and maximum permissible curvature mismatch between the prosthesis edge and the defect boundary.

[0012] Preferably, the prosthesis conflict index is a normalized risk measure that unifies constraints of different physical dimensions into a dimensionless scalar; the constraint evaluation unit is specifically used to calculate the actual geometric properties of each vertex on the initial prosthesis shape, calculate the risk ratio between the actual geometric properties and the surgical constraint limit, and take the maximum value among multiple risk ratios as the prosthesis conflict index.

[0013] Preferably, the conflict level classification process is as follows: when the prosthesis conflict index is less than or equal to 1, it is determined to be compliant; when the prosthesis conflict index is greater than 1 and less than or equal to 1.4, it is determined to be a level 1 conflict; when the prosthesis conflict index is greater than 1.4, it is determined to be a level 2 conflict.

[0014] Preferably, the specific process of the correction field generation unit is as follows: analyze the conflict vertices with a conflict index greater than 1, and deconstruct the normalized conflict index back into the corresponding physical violation quantity; generate a three-dimensional correction target vector pointing to the direction of the fastest gradient for the conflict vertices, forming a sparse correction target vector field; adopt a morphological correction propagation model based on Laplace deformation to transform the sparse correction target vector field into a dense displacement field to ensure a smooth transition between the corrected region and the uncorrected region.

[0015] Preferably, the closed-loop optimization unit obtains the corrected prosthetic shape by adding the vertex coordinate vector of the initial prosthetic shape to the dense displacement field; the system immediately uses the corrected prosthetic shape as the new input, calls the constraint evaluation unit again, and calculates the corrected conflict index.

[0016] Preferably, the graded final version strategy includes: when the maximum value of the corrected conflict index is less than or equal to 1.4, it is determined that the first-level correction has been achieved, the corrected prosthesis shape is subjected to final edge feathering processing, and a gradient pore structure is automatically generated in the non-load-bearing area; when the maximum value of the corrected conflict index is greater than 1.4, it is determined that the second-level conflict, the system automatically deploys a solid reinforcing rib structure in the corresponding area, keeps the area as a solid, and highlights the mark to prompt the operator to manually confirm.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] 1. This invention resolves the conflict between anatomical fidelity and clinical safety constraints in cranioplasty by constructing a system that incorporates morphological prediction, constraint assessment, correction generation, and closed-loop optimization. The system utilizes statistical shape models and patient CT data to generate highly personalized initial prostheses, ensuring anatomical accuracy in the repair. Compared to traditional mirror-image methods, this approach simultaneously guarantees automation efficiency and repair effectiveness when dealing with complex or asymmetric defects, providing clinicians with a high-quality repair solution that balances aesthetics and function.

[0019] 2. This invention innovatively establishes a multi-dimensional surgical constraint assessment mechanism, unifying and quantifying complex clinical requirements with different physical dimensions, such as prosthesis thickness, safe distance from key neurovascular structures, and edge curvature mismatch, into a dimensionless "prosthesis conflict index." This index provides a precise, unified, and operable quantitative basis for subsequent automated corrections. By classifying the conflict index, the system can clearly identify risk areas and levels, making subsequent optimization decisions more targeted and accurate, and significantly improving the safety of the design.

[0020] 3. This invention proposes an intelligent method for generating correction fields. The system can inversely analyze the quantified conflict index into specific physical violation quantities and generate sparse correction vectors. The key is that by introducing a propagation model based on Laplace deformation, the local, sparse correction target can be smoothly expanded into a dense displacement field covering the entire system. This ensures a smooth transition between the corrected and uncorrected regions, effectively avoiding the introduction of new geometric defects during the automatic correction process.

[0021] 4. The core advantage of this invention lies in its closed-loop optimization and tiered final version strategy of correction and verification. After the system applies a dense displacement field to correct the prosthesis, it immediately resubmits it to the constraint evaluation unit for verification, forming a decision-making closed loop. Based on the corrected conflict index, the system automatically executes a tiered final version: the compliant shape will undergo edge feathering and porous structure generation to achieve lightweighting; for secondary conflict areas, solid reinforcing ribs will be automatically deployed and highlighted for manual confirmation, maximizing design efficiency while ensuring absolute safety for clinical applications. Attached Figure Description

[0022] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0023] Figure 1 This is a system structure diagram of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0025] Example 1:

[0026] Please see Figure 1 A computer-aided cranial prosthesis design and manufacturing system includes a morphological prediction unit, a constraint evaluation unit, a correction field generation unit, and a closed-loop optimization unit.

[0027] The morphological prediction unit is used to solve for the optimal morphological coefficients and generate the initial prosthesis morphology based on a preset statistical shape model and the collected patient skull reference surface.

[0028] The constraint assessment unit is used to perform geometric analysis on the initial prosthesis shape based on preset multi-dimensional surgical constraint limits, calculate the prosthesis conflict index, and classify the conflict level according to the prosthesis conflict index.

[0029] The modified field generation unit is used to deconstruct the physical violation quantity in response to the prosthesis conflict index, and solve for the generation of a dense displacement field covering all vertices of the prosthesis through the morphological modification propagation model.

[0030] The closed-loop optimization unit is used to apply a dense displacement field to geometrically deform the initial prosthesis shape to obtain the corrected prosthesis shape, and then resubmits the corrected prosthesis shape to the constraint evaluation unit to execute the hierarchical final version strategy.

[0031] This embodiment provides a computer-aided cranial prosthesis design and manufacturing system. Through the collaborative work and closed-loop iteration of four major units, the system aims to solve the inherent conflicts between automation efficiency, anatomical fidelity and aesthetics, and clinical surgical safety constraints in existing technologies, such as the traditional mirror method, when dealing with complex or asymmetric defects.

[0032] The system includes:

[0033] Morphological prediction unit: The purpose of this unit is to generate a statistically optimal and highly personalized initial prosthesis base for the patient based on prior anatomical knowledge; in this embodiment, the unit requires two core inputs: a preset statistical shape model (SSM) and a collected reference surface of the patient's skull;

[0034] In this invention, the statistical shape model (SSM) refers to a pre-constructed prior model that describes the statistical regularities of skull morphology in a group. It consists of an average skull shape vector. and a set of principal mutation patterns, i.e., feature vectors The composition is obtained by training a large-scale standard skull dataset, i.e., sample size N, using morphological alignment and principal component analysis (PCA); the patient skull reference surface refers to the surface obtained from the patient's skull defect CT data. Automatically extracted remaining usable skull surface area excluding the defective region ;

[0035] When this unit runs, it is based on the input SSM ( ) and patient reference surface An optimal morphological coefficient vector is obtained through an optimization algorithm. Optimal morphological coefficients This refers to a set of scalar weights used to determine the main mutation patterns in SSM. A linear combination is performed to generate a reference surface that best fits the patient's own body. Complete skull morphology ;

[0036] The predictive shape generation process follows the standard linear combination equation of the statistical shape model:

[0037] ;

[0038] in, The complete skull morphology to be solved, conforming to statistical laws, is represented as a high-dimensional geometric position vector; unit: meters; the output of this formula.

[0039] : Population average skull shape vector; unit: meters; constant input provided in advance by SSM;

[0040] : No. Each principal variation pattern feature vector represents the main direction of morphological change; unit: meters; constant input provided in advance by SSM.

[0041] : Optimal morphological coefficient scalar weights, representing the individual's ... The standard deviation multiple of a mutation pattern; a dimensionless scalar; the objective unknown to be solved by a system optimization algorithm;

[0042] : Total number of principal patterns selected; Dimension: Integer; System constant preset for the number of patterns required to interpret 98% of the population variance according to SSM;

[0043] The system constructs the process as a constrained optimization problem, which is solved using an optimization algorithm that aims to minimize the predicted shape. Patient reference surface The fitting error between them is determined by regularization constraints. The value range of is, for example, within ±3 standard deviations of the mean, to avoid generating distorted shapes that do not conform to anatomical rules, and finally, the optimal value is obtained. Vector; the output of this process is complete. The system extracts the boundary of the missing part from it. This part is defined as the initial prosthesis morphology. ; It is passed down as the core processing object for all subsequent units;

[0044] Constraint Evaluation Unit: The purpose of this unit is to receive the initial prosthesis generated by the morphology prediction unit, which focuses on anatomical aesthetics. The unit performs quantitative assessments based on clinical safety guidelines; it also assesses surgical constraints based on pre-defined multi-dimensional limits. Geometric analysis is performed; multidimensional surgical constraint limits refer to a series of constant thresholds extracted from clinical surgical requirements and biomechanical specifications, such as minimum permissible strength thickness. Maximum appearance thickness limit Minimum safe distance from critical nerves or vascular sinuses wait;

[0045] This unit calculates through geometric analysis. Each vertex of the surface The actual geometric properties, such as the actual thickness at that point. Actual safe distance Actual curvature These measured values ​​were then compared with the aforementioned constraint limits to calculate the prosthesis conflict index. ; Prosthesis conflict index This invention defines a normalized risk measurement function, which aims to unify constraints of different physical dimensions onto a dimensionless scalar. Its calculation method is as follows:

[0046] ;

[0047] in, :vertex Conflict index at the location; Dimension: Dimensionless scalar; Calculation output of this formula;

[0048] : Prosthesis At the apex The measured geometric properties at the location, namely thickness, safety distance, and curvature; through the analysis of... The morphology is calculated in real time using geometric analysis.

[0049] : Preset multi-dimensional surgical constraint limits; system constants preset based on clinical guidelines or engineering experience;

[0050] Each internal factor represents a risk ratio; for example, when the implant is too thick ( When the implant is too thin, the first term is greater than 1; when the implant is too thin ( When ), the second term is greater than 1; Function ensures It captures the most serious violation at that point; if If the condition is met, then all constraints at that point are satisfied; any This indicates that at least one constraint is violated at that point;

[0051] The output of this unit is a scalar field, i.e., for each vertex. One The value is used to classify conflict levels, for example, Level 1 conflict: Second-level conflict: ;Should Scalar fields and conflict levels are passed to subsequent units;

[0052] Modified Field Generation Element: The purpose of this element is to generate a global geometric correction scheme that ensures a smooth transition in response to conflicts detected by the constraint evaluation element; this element receives... Scalar field, and analyze all conflicting vertices, i.e. The vertex; will be normalized Deconstructing back to its corresponding physical violation; the physical violation refers to the specific physical value of the conflict, such as the thickness violation. ,like or safe distance intrusion volume ,like ;

[0053] This unit is for each conflict vertex. Generate a three-dimensional corrected target vector Its direction points to The direction of the steepest gradient decreases, and the magnitude is positively correlated with the physical violation. Construct a sparse modified target vector field Finally, the unit employs a morphological correction propagation model, for example, a Laplace deformation-based model, which uses a sparse field. As a hard constraint, by solving the equation that minimizes the overall deformation energy of the mesh, the sparse correction objective is propagated to all vertices of the prototyping, and finally a dense displacement field covering all vertices of the prototyping is generated. ;

[0054] Closed-loop optimization unit: This unit is the core hub connecting prediction and correction, used to perform corrections, verify results, and execute the final strategy; this unit applies the dense displacement field generated in Section 3. The initial prosthesis morphology generated in the first section Perform geometric deformation; this process is achieved through standard three-dimensional vector addition:

[0055] ;

[0056] in, : Vertex coordinate vector of the corrected prosthetic shape; unit: meters; output of this formula;

[0057] : Vertex coordinate vector of the initial prosthesis shape; unit: meters; output of the shape prediction unit;

[0058] Dense displacement field; unit: meters; solution results for corrected field generation elements;

[0059] This calculation yields the corrected prosthesis morphology. The core technical feature of this unit is that the system obtains... Immediately afterwards, perform closed-loop verification: As new input, it is resubmitted to the constraint evaluation unit; the constraint evaluation unit uses the same method as described above. Formula pair A revised conflict index was obtained through recalculation. ;

[0060] The closed-loop optimization unit is based on this set The result, its maximum value To implement different graded final version strategies, for example, automatically deploying reinforcing ribs for areas that still have secondary conflicts after correction, and automatically generating pore structures for areas that have met the standards.

[0061] Technical Effects: This invention constructs a complete prediction-evaluation-correction-verification technical closed loop by organically combining a morphological prediction unit (which addresses aesthetics and fidelity), a constraint evaluation unit, a quantification of safety conflicts, a correction field generation unit, a calculation of smoothing correction schemes, and a closed-loop optimization unit, to perform corrections, verify results, and execute the final version strategy. This system can automatically resolve the conflict between anatomical high fidelity and clinical surgical safety constraints, ensuring that the final prosthesis design meets both requirements simultaneously, significantly improving the design efficiency and safety of complex skull defect repair.

[0062] Example 2:

[0063] The statistical shape model includes the average skull shape vector and principal variation pattern constructed by morphological alignment and principal component analysis of a large-scale standard skull dataset; the patient skull reference surface is obtained by acquiring CT data of the patient's skull defect and automatically extracting the defect boundary and the remaining effective skull surface;

[0064] This embodiment, based on embodiment 1, specifies the two core input data sources upon which the morphology prediction unit depends;

[0065] The Statistical Shape Model (SSM), serving as a systematic anatomical prior knowledge base, is constructed as follows: A large-scale, e.g., N-sample, standard skull dataset is collected; morphological alignment is performed on this dataset to ensure all samples are in a uniform coordinate and pose space, followed by Principal Component Analysis (PCA); the results of PCA are refined and stored into two core components: the average skull shape vector, i.e. This represents the average morphology and a set of main variation patterns of the population, namely... The vector group represents the main statistical directions of group morphological differences; these two components together constitute the SSM input required by the morphological prediction unit.

[0066] The patient's skull reference surface is used as the basis for personalized matching by the system. The acquisition process is as follows: the system collects CT data of the patient's skull defects, i.e. Through image segmentation and geometric processing algorithms, the system automatically selects from... Two key pieces of information were extracted: one was the defect boundary of the defect area, i.e. Secondly, the remaining effective skull surface, i.e. ;this It is used as a reference surface for fitting calculations with SSM;

[0067] By constructing the SSM using the PCA method, this invention obtains a highly condensed and statistically significant anatomical prior model. This ensures that the anatomical basis for morphological prediction is robust and standardized; simultaneously, by automatically extracting from patient CT data... As a reference surface, this ensures that the final result of morphological prediction is highly personalized and accurately matches the remaining bone structure of the specific patient, thus solving for the optimal coefficients for the morphological prediction unit. It provides precise input.

[0068] Example 3:

[0069] Multidimensional surgical constraint limits include: minimum permissible strength thickness of the prosthesis, maximum appearance limitation thickness, minimum safe distance from key nerves or vascular sinuses, and maximum permissible curvature mismatch between the prosthesis edge and the defect boundary;

[0070] This embodiment, based on embodiment 1, further defines the specific content of the multi-dimensional surgical constraint limits upon which the constraint evaluation unit is based; these limits are key parameters for ensuring the clinical safety and engineering effectiveness of the prosthesis, and are pre-set by clinical guidelines or engineering experience and used as constant input systems;

[0071] In this embodiment, these limits specifically include, but are not limited to, the following four dimensions:

[0072] Minimum allowable strength thickness of prosthesis ( Biomechanical specifications, such as a setting of 3mm, aim to ensure that the prosthesis has sufficient mechanical strength at any point to withstand impact and prevent repair failure.

[0073] Maximum appearance thickness ( ): Clinical aesthetic requirements, for example, set to 8mm; the purpose is to prevent the prosthesis from being too thick in certain areas (especially the edge fitting part of non-defect areas), resulting in postoperative bulging or excessive flap tension;

[0074] Minimum safe distance from critical nerves or vascular sinuses ( Anatomical specifications; the purpose is to ensure that any part of the prosthesis, especially the inner surface, maintains an absolutely safe gap with critical intracranial structures such as the superior sagittal sinus or important neural pathways to prevent compression or damage.

[0075] The maximum permissible curvature mismatch between the prosthesis edge and the defect boundary ( The purpose is to ensure that the connection between the prosthesis edge and the patient's own bone window edge is geometrically smooth and continuous, avoiding steps or sharp transitions, which is crucial for stress distribution and edge fit.

[0076] By clearly defining the intensity of this coverage ( ), aesthetically pleasing ),Safety( ) and edge matching ( The invention provides a comprehensive and accurate quantitative assessment standard for constraint assessment units by addressing multi-dimensional physical constraints; this enables the subsequently calculated prosthesis conflict index to accurately reflect all potential clinical risks.

[0077] Example 4:

[0078] The prosthesis conflict index is a normalized risk measure that unifies constraints of different physical dimensions into a dimensionless scalar. The constraint assessment unit is specifically used to calculate the actual geometric properties of each vertex on the initial prosthesis shape, and to calculate the risk ratio between the actual geometric properties and the surgical constraint limit, and to take the maximum value among multiple risk ratios as the prosthesis conflict index.

[0079] This embodiment, based on Embodiment 1, calculates the prosthesis conflict index using the constraint evaluation unit. A detailed explanation of the specific methods and technical aspects;

[0080] Prosthesis Conflict Index In this invention, it is designed as a normalized risk measure, comprising multiple quantities with different physical dimensions, such as thickness. The dimensions of are length and curvature. The dimensionless constraint of 1 / length is unified into a dimensionless scalar. This is to facilitate unified evaluation and comparison;

[0081] The specific calculation process for the constraint evaluation unit is as follows:

[0082] The system receives the initial prosthesis shape. And calculate the vertices of it. The actual geometric properties, such as ;

[0083] Next, the system calculates these actual attributes against preset surgical constraint limits, such as... The risk ratio refers to the dimensionless ratio obtained by comparing the measured value with the limit. Its design ensures that any violation, such as being too thick, too thin, or too close, will result in a ratio greater than 1. In this embodiment, these risk ratios are specifically:

[0084] That is, assess the risk of excessive thickness;

[0085] That is, assess the risk of it being too thin;

[0086] That is, assess the risk of being too close;

[0087] That is, assess the risk of curvature mismatch;

[0088] To ensure robustness of the calculation, the system performs a validity check before performing the division operation. This is especially important in cases where the denominator might be zero or a very small value, such as measured thickness. or safe distance The system will directly set the corresponding risk ratio to a preset maximum value, or determine it as the highest conflict level, in order to avoid calculation errors and ensure that such extreme risk points can be effectively captured and handled.

[0089] The system takes the maximum value among the above multiple risk ratios as the vertex. Final prosthesis conflict index ;

[0090] By defining the conflict index as the maximum value of multiple risk ratios, this invention ensures... It can sensitively capture the most serious violation at a given point; this design unifies and quantifies all different types of risk into a single scalar, greatly simplifying the input of subsequent correction units; ensuring that as long as This point must satisfy all constraints, and any All points must be conflict points that need to be corrected, providing accurate, reliable and easy-to-process quantitative basis for subsequent conflict classification and correction field generation.

[0091] Example 5:

[0092] The conflict level classification process is as follows: when the prosthesis conflict index is less than or equal to 1, it is judged as compliant; when the prosthesis conflict index is greater than 1 and less than or equal to 1.4, it is judged as Level 1 conflict; when the prosthesis conflict index is greater than 1.4, it is judged as Level 2 conflict.

[0093] This embodiment, based on embodiment 1, calculates the constraint evaluation unit... After determining the value, the specific constraints on how to implement the conflict level classification should be defined; this classification process forms the basis for subsequent decisions regarding the final version of the classification strategy.

[0094] The system calculates the prosthesis conflict index based on Example 4. For each vertex, the following fixed threshold rule is applied for classification:

[0095] Compliance: When the prosthesis conflict index is less than or equal to 1, i.e. When the vertex is deemed compliant, it indicates that the vertex matches the statistical model in terms of anatomical morphology and simultaneously satisfies all the multidimensional surgical constraints defined in Example 3.

[0096] Level 1 conflict: When the prosthesis conflict index is greater than 1 and less than or equal to 1.4, i.e. When the vertex's shape violates at least one constraint, it is considered a Level 1 conflict. This indicates that the shape of the vertex violates at least one constraint, constituting a Level 1 risk. This conflict is considered to be resolved through automatic shape correction by the system, such as Laplace deformation. The threshold of 1.4 here is an empirical dividing line used in the field to distinguish between general risks and severe risks. For example, it can be obtained by statistical analysis of risk data from historical repair cases and is used to define the upper limit of conflicts that can be resolved through automatic smoothing correction.

[0097] Second-degree conflict: When the prosthesis conflict index is greater than 1.4, i.e. When the vertex's shape seriously violates the constraints, or there is a fundamental conflict between the statistically predicted shape and the surgical safety constraints that is difficult to reconcile through smooth deformation, thus constituting a level two risk;

[0098] By setting two clear numerical thresholds based on clinical risk statistics—the 1.0 compliance / non-compliance boundary and the 1.4 Level 1 / Level 2 risk boundary—this invention continuously... The risk scalar field is transformed into a discrete conflict level diagram with clear handling directions; this classification provides a clear execution path for subsequent closed-loop optimization units: that is, distinguishing which conflicts should be automatically smoothed and corrected (Level 1), and which conflicts represent more serious problems and need to be supplemented with additional structural reinforcement and manual confirmation after automatic correction (Level 2).

[0099] Example 6:

[0100] The specific process of generating the correction field unit is as follows: Analyze the conflict vertices with a conflict index greater than 1, and deconstruct the normalized conflict index back into the corresponding physical violation quantity; generate a three-dimensional correction target vector pointing to the direction of the steepest gradient for the conflict vertices, forming a sparse correction target vector field; adopt a morphological correction propagation model based on Laplace deformation to transform the sparse correction target vector field into a dense displacement field to ensure a smooth transition between the correction region and the uncorrected region.

[0101] Based on Example 1, this embodiment details how the field generation unit intelligently generates a dense displacement field from the conflict index, i.e., the evaluation result, i.e., the specific technical process of the correction scheme.

[0102] The specific process of this unit is divided into three logical steps:

[0103] Conflict deconstruction: The system analyzes all conflict vertices with a conflict index greater than 1, i.e., the first and second-level conflict points defined in Example 5; for these points, the system performs inverse calculations to normalize the dimensionless conflict index. Deconstruct it back to its corresponding physical violation; for example, if a point It is by The system then calculates the specific thickness deviation. Unit: meters;

[0104] Sparse field construction: Based on the physical violation dataset obtained in step 1, the system constructs a sparse field for each conflict vertex. Generate a three-dimensional corrected target vector pointing in the direction of the steepest gradient. The steepest gradient direction refers to the direction that reduces the gradient the fastest. The geometric direction of the value, for example, for excessively thin points, points outward from the surface normal to increase thickness; for excessively close points, it points away from the hazard source; this vector The modulus and the corresponding physical violation, such as Positive correlation; all these discrete The vectors together constitute the sparse modified target vector field. The area is defined only at the point of conflict;

[0105] Propagation and Solution: Due to sparse fields Defining only a local correction target and applying it directly can lead to steps or sharp points on the prosthesis surface. To address this issue, this embodiment employs a morphological correction propagation model based on Laplace deformation. This model is a mature mesh deformation technique that utilizes sparse fields... As a hard constraint, the positions of these points must satisfy the following after correction: The defined objective is to minimize the overall deformation energy of the mesh while maintaining shape smoothness. This ensures that local geometric corrections diffuse smoothly and uniformly across the entire prosthesis surface, much like heat conduction. The model propagates local correction requirements to all vertices of the prosthesis, ultimately calculating a globally optimal and uniformly smooth dense displacement field. ;

[0106] This embodiment achieves an intelligent correction scheme generation process through three steps: deconstruction, sparse field, and propagation. In particular, by employing a propagation model based on Laplace deformation, this invention transforms the local, discrete, and potentially conflicting surgical constraint correction target, i.e., the sparse field C, into a global, continuous, and smooth geometric deformation scheme dense displacement field. This ensures that the automatic correction process smoothly transitions between the corrected and uncorrected areas, avoiding the introduction of new geometric defects, such as steps or sharp edges, by the automatic correction itself.

[0107] Example 7:

[0108] The closed-loop optimization unit obtains the corrected prosthetic shape by adding the vertex coordinate vector of the initial prosthetic shape to the dense displacement field. The system immediately uses the corrected prosthetic shape as the new input, calls the constraint evaluation unit again, and calculates the corrected conflict index.

[0109] This embodiment, based on embodiment 1, specifies the specific implementation method of the core closed-loop operation of the closed-loop optimization unit performing correction-verification;

[0110] The closed-loop operation of this unit consists of two closely linked actions:

[0111] Correction Execution: The closed-loop optimization unit performs geometric deformation; this process involves modifying the vertex coordinate vectors of the initial prosthetic shape, i.e. The dense displacement field generated by the morphology prediction unit and the embodiment is... The modified scheme involves vector addition to obtain the corrected prosthesis shape. The mathematical expression of this process is... ;

[0112] Closed-loop verification: during generation Then, the system immediately performs a closed-loop verification step, which involves verifying the corrected prosthesis shape. As new input, the constraint evaluation unit is invoked again; the constraint evaluation unit, following the method of Example 4, evaluates... A completely new geometric analysis is performed on each vertex, and a new set of corrected conflict indices is calculated, namely... ;

[0113] This embodiment constructs a complete correction-verification feedback loop through two steps: applying correction via vector addition and immediately re-evaluating. This design ensures that the system does not simply blindly execute a correction scheme. ), and even more so the result of the correction ( To obtain the corrected actual conflict index, immediate, quantitative quality control checks are performed. ); This is verified This provides a final and reliable basis for decision-making regarding subsequent tiered final versions of the strategy.

[0114] Example 8:

[0115] The graded final version strategy includes: when the maximum value of the corrected conflict index is less than or equal to 1.4, it is judged as a first-level correction that meets the standard. The corrected prosthesis shape is then subjected to final edge feathering, and a gradient porosity structure is automatically generated in the non-load-bearing area. When the maximum value of the corrected conflict index is greater than 1.4, it is judged as a second-level conflict. The system automatically deploys a solid reinforcing rib structure in the corresponding area, keeps the area as a solid, and highlights it to prompt the operator to manually confirm.

[0116] This embodiment, based on Embodiment 1, specifies the hierarchical final version strategy executed by the closed-loop optimization unit after completing the closed-loop verification in Embodiment 7. This strategy is the system's final automated decision-making exit, and it is based on the modified maximum conflict index, i.e. The value automatically selects different manufacturing data generation paths;

[0117] The tiered final version strategy includes two scenarios:

[0118] Level 1 correction automatically meets the standard

[0119] Triggering condition: When the maximum value of the corrected conflict index is less than or equal to 1.4, i.e. Triggered at time;

[0120] Judgment: This condition indicates that, after closed-loop correction, all remaining conflicts, whether Level 1 or Level 2, have been reduced to below the Level 1 conflict threshold (1.4) or are fully compliant. The system has determined that the automatic correction has met the requirements.

[0121] Action performed: The system modifies the prosthesis shape. The procedure includes: final edge feathering, such as a 0.5mm chamfer, to ensure edge fit; and in non-load-bearing areas, such as areas with a thickness greater than 6mm, based on mechanical analysis or thickness determination, automatically generating a gradient porosity structure, such as 20%-40% porosity. The purpose is to: significantly reduce the weight of the prosthesis and reduce pressure on the underlying soft tissue; provide a biological scaffold for the inward growth of new bone tissue, promote the bony integration of the prosthesis and autologous bone, thereby enhancing the long-term stability and biocompatibility of the implant.

[0122] Secondary corrections require manual intervention.

[0123] Triggering condition: When the maximum value of the corrected conflict index is greater than 1.4, i.e. Triggered at time;

[0124] Judgment: This condition indicates the existence of a secondary conflict zone, that is, a high degree of conflict between the statistically predicted aesthetic morphology and clinical safety constraints, such as minimum thickness or safety distance, in this zone, and even Laplace smoothing deformation failed to reduce its risk to below 1.4.

[0125] Execution action: The system automatically performs the action in the corresponding area, i.e. The vertex and its neighborhood, such as a 3-5mm range: deploy solid reinforcing rib structures, such as 1.5mm thick reinforcing ribs, to force mechanical requirements to be met without severely damaging the appearance; keep this area as solid, i.e. set to 0% porosity; and highlight this area to prompt the operator to make a final manual confirmation to ensure that the operator notices the automatic processing results of this high-risk area.

[0126] Through this tiered final version strategy, the present invention achieves an intelligent decision-making closed loop. For most conflicts that can be resolved by smooth deformation, the system realizes a fully automated process for generating lightweight porous prostheses. For rare areas where there is an extreme conflict between statistical prediction and clinical safety, the system automatically switches to a safety-first strategy, automatically adding reinforcing ribs to ensure structural safety and mandating manual review. This ensures a very high level of automation while guaranteeing clinical safety in extreme cases.

[0127] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A computer-aided based cranial prosthesis design and manufacturing system, characterized by, It includes a morphology prediction unit, a constraint evaluation unit, a correction field generation unit, and a closed-loop optimization unit; The morphological prediction unit is used to solve for the optimal morphological coefficients and generate the initial prosthesis morphology based on a preset statistical shape model and the collected patient skull reference surface. The constraint assessment unit is used to perform geometric analysis on the initial prosthesis shape based on preset multi-dimensional surgical constraint limits, calculate the prosthesis conflict index, and classify the conflict level according to the prosthesis conflict index. The modified field generation unit is used to deconstruct the physical violation quantity in response to the prosthesis conflict index, and solve for the generation of a dense displacement field covering all vertices of the prosthesis through the morphological modification propagation model. The closed-loop optimization unit is used to apply a dense displacement field to geometrically deform the initial prosthesis shape to obtain the corrected prosthesis shape, and then resubmits the corrected prosthesis shape to the constraint evaluation unit to execute the hierarchical final version strategy. The specific process of the modified field generation unit is as follows: Analyze the conflict vertices with a prosthesis conflict index greater than 1, deconstruct the normalized conflict index back into the corresponding physical violation quantity; generate a three-dimensional modified target vector pointing to the direction of the steepest gradient for the conflict vertices, and form a sparse modified target vector field. A Laplace deformation-based morphological correction propagation model is adopted to transform the sparse correction target vector field into a dense displacement field, so as to ensure a smooth transition between the corrected and uncorrected regions.

2. The computer-aided cranial prosthesis design and manufacturing system according to claim 1, characterized in that, The statistical shape model includes the average skull shape vector and principal variation pattern constructed by morphological alignment and principal component analysis of a large-scale standard skull dataset; the patient skull reference surface is obtained by acquiring CT data of the patient's skull defect and automatically extracting the defect boundary and the remaining effective skull surface.

3. The computer-aided cranial prosthesis design and manufacturing system according to claim 1, characterized in that, The multidimensional surgical constraint limits include: minimum permissible strength thickness of the prosthesis, maximum appearance restriction thickness, minimum safe distance from critical nerves or vascular sinuses, and maximum permissible curvature mismatch between the prosthesis edge and the defect boundary.

4. The computer-aided based cranial prosthesis design and manufacturing system of claim 1, wherein, The prosthesis conflict index is a normalized risk measure that unifies constraints of different physical dimensions into a dimensionless scalar. The constraint assessment unit is specifically used to calculate the actual geometric properties of each vertex on the initial prosthesis shape, calculate the risk ratio between the actual geometric properties and the surgical constraint limit, and take the maximum value among multiple risk ratios as the prosthesis conflict index.

5. The computer-aided based cranial prosthesis design and manufacturing system of claim 1, wherein, The conflict level classification process is as follows: when the prosthesis conflict index is less than or equal to 1, it is judged as compliant; when the prosthesis conflict index is greater than 1 and less than or equal to 1.4, it is judged as a level 1 conflict; when the prosthesis conflict index is greater than 1.4, it is judged as a level 2 conflict.

6. A computer-aided based cranial prosthesis design and manufacturing system according to claim 1, wherein, The closed-loop optimization unit obtains the corrected prosthetic shape by adding the vertex coordinate vector of the initial prosthetic shape to the dense displacement field. The system immediately uses the corrected prosthetic shape as the new input, calls the constraint evaluation unit again, and calculates the corrected conflict index.

7. The computer-aided cranial prosthesis design and manufacturing system according to claim 1, characterized in that, The graded final version strategy includes: when the maximum value of the corrected conflict index is less than or equal to 1.4, it is determined to be a first-level correction that meets the standard. The corrected prosthesis shape is then subjected to final edge feathering processing, and a gradient pore structure is automatically generated in the non-load-bearing area. When the maximum value of the corrected conflict index is greater than 1.4, it is determined to be a second-level conflict. The system automatically deploys a solid reinforcing rib structure in the corresponding area, keeps the area as a solid, and highlights it to prompt the operator to manually confirm.

Images