Cross-modality registered denture design method

Abstract

The application provides a removable partial denture design method for cross-modality registration, comprising the following steps: acquiring two-dimensional oral panoramic film and three-dimensional intraoral scanning data of a patient; calculating the radiological bone loss proportion of each tooth, and obtaining the image reliability component of each tooth position in the panoramic film based on the segmentation result; extracting the three-dimensional key point and neck margin contour point set of each tooth; calculating the registration reliability component of each tooth position; fusing the image reliability component and the registration reliability component of each tooth position in the panoramic film to generate an image and registration reliability index; and taking the radiological bone loss proportion and the reliability index of each tooth position as a design constraint to perform automatic design of the removable partial denture. The purpose of improving the clinical scientificity, safety and interpretability of the design scheme is achieved.

Description

Technical Field

[0001] This invention belongs to the field of digital technology of oral restoration, and relates to medical image processing and three-dimensional oral model calculation technology. In particular, it relates to a method for quantitatively evaluating radiographic bone loss information based on cross-modal registration of panoramic oral radiographs and intraoral scan data, and using it to assist in the design of removable partial dentures. Background Technology

[0002] The design of removable partial dentures needs to consider the following factors simultaneously: the extent of missing teeth (Kennedy classification), the health status of remaining teeth (such as mobility, tilt angle, restoration, caries status, etc.), periodontal support (degree of alveolar bone resorption), soft tissue morphology (mucosal active area), and the denture's path of insertion and undercuts. Among these, alveolar bone height and the degree of loss are important criteria for assessing the stability of remaining teeth.

[0003] Radiographic bone loss (RBL) is a commonly used imaging assessment indicator, typically based on the ratio of alveolar bone height to tooth root length in panoramic dental radiographs, used to assess the severity of periodontal disease and the support of periodontal tissues. In recent years, studies have attempted to use deep learning methods to automatically assess RBL from panoramic dental radiographs to aid in the early identification of periodontal disease. However, the panoramic dental radiograph imaging process is highly technically sensitive and easily affected by factors such as equipment parameters, patient positioning, and overlapping anatomical structures, leading to image distortion and measurement instability.

[0004] The existing technology has the following shortcomings:

[0005] (1) Existing RBL assessment methods based on panoramic radiographs are mainly for diagnostic or screening applications, and are usually output in the form of overall or graded results of the dentition. They cannot provide calculable structured parameters that can be directly used for denture design at the tooth-by-tooth level.

[0006] (2) In the process of digital denture design, intraoral scanning data can accurately reflect the three-dimensional morphology of the crown and soft tissue, but it cannot provide radiographic information such as the height of the tooth root and alveolar bone. There is a lack of a reliable mechanism in the existing technology to correspond and map the RBL information obtained from panoramic radiographs with the three-dimensional tooth position entities obtained from intraoral scanning;

[0007] (3) Due to differences in image quality and cross-modal registration errors, directly using panoramic radiographs for denture design may lead to unreasonable selection of abutment teeth or setting of retention strength. Existing technologies have not yet explicitly quantified the unreliability of the above-mentioned images and registration process and incorporated it into the denture design decision-making process.

[0008] Therefore, it is necessary to propose a technical solution that can assess radiographic bone loss in panoramic radiographs, quantify image quality and cross-modal registration reliability, and use the relevant results to assist in the automated design of removable partial dentures. Summary of the Invention

[0009] To address the problems existing in the prior art, this invention provides a method for designing removable partial dentures with cross-modal registration. This method evaluates radiographic bone loss information at each tooth position in panoramic radiographs and introduces an image-registration reliability index to quantify image quality and cross-modal mapping reliability. This transforms relevant image information into computable design constraints to assist in the automated design of removable partial dentures.

[0010] This disclosure provides a method for designing removable partial dentures with cross-modal registration, including:

[0011] Acquire two-dimensional panoramic radiographs and three-dimensional intraoral scans of the patient;

[0012] The panoramic film is subjected to tooth-by-tooth detection and segmentation. Based on the segmentation results, the reference point of the cementoenamel junction, the apex point, and the mesial and distal ridge apex of each tooth are extracted. The radiographic bone loss ratio of each tooth is calculated based on the reference point, the apex point, and the ridge apex. Based on the segmentation results, the image confidence component of each tooth position in the panoramic film is obtained.

[0013] The intraoral scan data is segmented into tooth structure and gingiva to construct a three-dimensional tooth position entity, and the three-dimensional key points and cervical margin contour point set of each tooth are extracted.

[0014] A parametric projection model of the dental arch is established based on the three-dimensional key points and cervical margin contour point set of each tooth. Cross-modal registration is performed on the panoramic radiograph and intraoral scan data. The registration residual is obtained by solving the projection model parameters through optimization algorithm, and the registration confidence component of each tooth position is calculated based on the registration residual.

[0015] The image confidence component of each tooth position in the panoramic film is fused with the registration confidence component to generate an image and registration confidence index.

[0016] The proportion of radiographic bone loss and the reliability index of each tooth position are used as design constraints to automate the design of removable partial dentures.

[0017] Optionally, the calculation of the radiographic bone loss percentage for each tooth includes:

[0018] Based on formula Calculate the root length of the tooth;

[0019] Based on formula Calculate near-mid height based on the formula Calculate the distance to mid-distance;

[0020] The proportions of mesial and distal bone loss are calculated based on the root length, mesial height, and distal height of the tooth position.

[0021] The larger of the mesial and distal bone loss proportions is selected as the radiographic bone loss proportion of the tooth;

[0022] in, For root length, For the root tip, As a reference point for the cementoenamel junction, Near to medium altitude, At a distance of mid-altitude, To project onto the unit vector of the major axis, It is the proximal bony ridge. It is the distal ridge of the bone.

[0023] Optionally, the factor for calculating the image confidence component in the image confidence component of each tooth position in the panoramic film based on the segmentation results includes:

[0024] Confidence of tooth segmentation, clarity of tooth margin gradient, and overlap score of adjacent tooth masking;

[0025] The formula for calculating the image credibility component is as follows: ,in, For image credibility components, The confidence level for tooth position segmentation. For the clarity of the gradient at the tooth margin, The overlap score of adjacent tooth position masks is calculated.

[0026] Optionally, the step of establishing a parametric projection model of the dental arch based on the three-dimensional key points and cervical margin contour point set of each tooth includes:

[0027] 3D points Using formula Perform a rigid body transformation, and then convert the transformation result... Mapped to panoramic pixel coordinates ;

[0028] The mapping function is:

[0029] ,

[0030] ,

[0031] in and For point The corresponding dental arch parameter coordinates, For arc length parameters, For height parameters, Let be a rotation matrix. It is a translation vector. This is the offset. The coefficient of the linear term, The coefficient of the quadratic term, As the vertical baseline, This is the vertical scaling factor. For coupling term coefficients, parameter set .

[0032] Optionally, the objective function for obtaining the registration residual by solving the projection model parameters through an optimization algorithm includes:

[0033] ,

[0034] in, , and All are weighting coefficients. For key point error, For contour error, As a constraint used to maintain the monotonicity of tooth position sequence;

[0035] ,

[0036] ,

[0037] ,

[0038] in, The total error of the keypoint matching items. This represents the total error of the contour matching item. This is the total error of the constraint term used to maintain the monotonicity of tooth position order. For Huber's robustness loss, For collection points, For the minimum interval, For the first The coordinates of the two-dimensional key points corresponding to each tooth on a two-dimensional panoramic image. For the parameterized projection function of the dental arch, For the first The coordinates of the three-dimensional key points corresponding to each tooth on the three-dimensional intraoral scan model. For the first The signed distance field function of a tooth on a two-dimensional panoramic film. For the first The coordinates of the three-dimensional key points corresponding to each tooth on the three-dimensional intraoral scan model. These are the horizontal coordinates of the two-dimensional point after projection.

[0039] Optionally, the formula for calculating the credibility index is:

[0040] ,

[0041] in, As a credibility indicator, This is a quantification of the unreliability at the image level. This is the quantification result of the unreliability at the cross-modal registration level. The larger the value, the lower the reliability of the RBL result corresponding to that tooth position.

[0042] Optionally, the role of the image and registration reliability index includes:

[0043] Risk penalty factor used in abutment tooth suitability scoring;

[0044] Constraint factors used for calculating the upper limit of the depth for undercutting;

[0045] Criteria used for the selection and combination restrictions of fixation body types;

[0046] Used as the basis for risk warnings and review markings in electronic design sheets.

[0047] Optionally, the formula for calculating the suitability score of the abutment teeth is as follows:

[0048] ,

[0049] in, For radiographic indicators of bone loss, As a credibility indicator, The normalized result for looseness, This is the normalized result for the tilt angle. This serves as a risk marker for medical records.

[0050] Optionally, the constraint factor calculated using the upper limit of the indentation depth is:

[0051] .

[0052] Optionally, after the step of automating the design of removable partial dentures using the radiographic bone loss ratio and reliability index of each tooth position as design constraints, the method further includes:

[0053] The steps involve verifying and automatically repairing the automated design results based on manufacturing rules and clinical safety to obtain an electronic design sheet and manufacturable design data for removable partial dentures.

[0054] The present invention provides a cross-modal registration method for designing removable partial dentures. This method automatically assesses radiographic bone loss (RBL) at each tooth position in panoramic radiographs and maps it to a 3D dental arch scan through cross-modal registration. Simultaneously, it introduces the Image and Registration Reliability Index (CRI) to quantify the reliability of the RBL results. Using RBL and CRI as constraints, it automatically generates safe and interpretable removable partial denture design schemes and electronic design sheets. This accurately and reliably integrates key information from two-dimensional images into the three-dimensional digital design process. Under the premise of quantitatively assessing data reliability, it achieves personalized intelligent design of removable partial dentures driven by objective image data, with known and controllable risks, thereby improving the clinical scientificity, safety, and interpretability of the design scheme. Attached Figure Description

[0055] The above and other objects, features and advantages of this disclosure will become more apparent from the accompanying drawings, in which like reference numerals generally denote like parts.

[0056] Figure 1 A flowchart illustrating the cross-modal registration removable partial denture design method provided in this disclosure. Detailed Implementation

[0057] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0058] It should be understood that the following specific examples illustrate the implementation of this disclosure, and those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0059] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0060] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this disclosure. The illustrations only show the components related to this disclosure and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0061] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0062] RBL (Radiographic Bone Loss) refers to radiographic bone loss.

[0063] CEJ (Cemento-Enamel Junction) refers to the cementoenamel junction;

[0064] CRI (Confidence / Reliability Index) refers to the confidence index of image and registration.

[0065] Tooth position: FDI coding is used;

[0066] Location: including mesial / distal / buccal / lingual.

[0067] Radiographic bone loss, including impaction of teeth ,definition:

[0068] Root length : The distance from the CEJ reference point to the apex along the tooth's long axis.

[0069] To characterize the reliability of tooth-by-tooth RBL results in engineering and clinical applications, an image and registration reliability index was introduced. ∈[0,1], the larger the value, the lower the credibility.

[0070] CRI consists of the following two parts:

[0071] Image credibility component : Reflects the image quality and measurement stability of this tooth position in the panoramic radiograph;

[0072] Registration confidence components : Reflects the local reliability of 2D–3D cross-modal mapping.

[0073] Overall credibility is defined as:

[0074]

[0075] This definition is used for risk union modeling to ensure that the overall design strategy automatically tends to be conservative when any source is unreliable.

[0076] The technical objectives to be achieved in this embodiment include:

[0077] 1. How to automatically obtain a calculable RBL index for each tooth in a panoramic radiograph;

[0078] 2. How to establish cross-modal registration mapping between 2D panoramic images and 3D scanning images, and how to quantify the registration quality;

[0079] 3. How to unify the expression of image measurement error and registration error into a per-tooth position reliability index;

[0080] 4. How to transform RBL and its credibility into calculable constraints in the design of removable partial dentures (abutment tooth selection, upper limit of undercut utilization depth, clasp type restrictions, etc.).

[0081] This embodiment can automatically evaluate RBL in panoramic radiographs and, with its credibility clearly defined, reliably map the results to 3D tooth structures to drive the design of removable partial dentures.

[0082] Specifically, such as Figure 1 As shown, this embodiment discloses a method for designing removable partial dentures with cross-modal registration, characterized by comprising:

[0083] Step S101: Acquire the patient's two-dimensional panoramic oral radiographs and three-dimensional intraoral scan data;

[0084] Step S102: Perform tooth-by-tooth detection and segmentation on the panoramic film, and extract the cementoenamel junction reference point, apex point and mesial and distal ridge apex of each tooth based on the segmentation results. Calculate the radiographic bone loss ratio of each tooth based on the reference point, apex point and ridge apex, and obtain the image confidence component of each tooth position in the panoramic film based on the segmentation results.

[0085] Step S103: The intraoral scan data is segmented into tooth and gingival segments to construct a three-dimensional tooth position entity, and the three-dimensional key points and cervical margin contour point set of each tooth are extracted.

[0086] Step S104: Establish a parametric projection model of the dental arch based on the three-dimensional key points and cervical margin contour point set of each tooth, perform cross-modal registration of the panoramic radiograph and intraoral scan data, obtain the registration residual by solving the projection model parameters through optimization algorithm, and calculate the registration confidence component of each tooth position based on the registration residual.

[0087] Step S105: Fuse the image confidence component of each tooth position in the panoramic film with the registration confidence component to generate an image and registration confidence index;

[0088] Step S106: Using the radiographic bone loss ratio and reliability index of each tooth position as design constraints, perform automated design of removable partial dentures.

[0089] In step S101, the panoramic image P is read and grayscale normalization and distortion / noise preprocessing are performed (such as median filtering and CLAHE enhancement).

[0090] Read the scanning S, and perform noise reduction, hole repair, and resampling on the mesh (target side length such as 0.2–0.5 mm).

[0091] If there is a medical record E, it will be uniformly mapped to the FDI tooth position index and its consistency will be verified (for example, if the medical record is missing but the dental crown is detected by the oral scan, it will be marked as "conflict pending verification").

[0092] Step S102, which involves tooth-by-tooth feature extraction and tooth-by-tooth RBL information calculation from panoramic radiographs, includes:

[0093] Step S1021, Tooth position detection and segmentation:

[0094] Use any of the following implementations:

[0095] Deep Learning: Using U-Net / Mask R-CNN to output pixel masks for each tooth With confidence level .

[0096] Alternatively, a combination of threshold segmentation, morphological analysis, and connected component segmentation can be used, along with tooth row order and width constraints, to obtain the desired results. .

[0097] Step S1022: Extraction of root apex, CEJ reference point, and bony ridge point:

[0098] For each tooth mask :

[0099] Tooth major axis estimation: Perform PCA on the mask pixels and take the first principal direction as the tooth major axis direction. ;

[0100] Root tip :along The direction is determined by selecting the farthest boundary point of the mask as the candidate root tip. If a double-root morphology exists, the geometric center of the two root tips or the deepest root tip is selected as the root tip. .

[0101] CEJ Reference Point Draw equidistant cross-sections along the long axis of the tooth, calculate the length of the intersection between each cross-section and the mask, and take the position where the intersection length changes the most rapidly from large to small as the position of the narrow neck of the crown-root transition. Take the center point of the mask on this section line as... .

[0102] mesial / distal bony ridges Define a Region of Interest (ROI) on both the mesial and distal sides of the tooth (the area adjacent to the interdental space, within 2–6 mm of the tooth boundary); perform edge detection within the ROI to obtain candidate bone ridge curves;

[0103] Choose the option that satisfies: located at and Between, and along perpendicular to The point with the largest local gradient intensity is taken as the bone ridge; respectively... .

[0104] Distance calculation:

[0105] Root length ;

[0106] Near-to-medium height (Projected onto the major axis unit vector) );

[0107] Far-to-mid altitude ;

[0108] Proportion of mesial bone loss: ;

[0109] Distal bone loss percentage: ;

[0110] Overall RBL: ,in ∈[0,1], the larger the value, the more severe the degree of bone loss.

[0111] Step S1023, Image Confidence Component :

[0112] Segmentation confidence Confidence level The output can be obtained directly using ke; alternatively, it can be obtained by fitting edge sharpness.

[0113] Edge sharpness The mean gradient magnitude within the ROI is normalized to [0,1].

[0114] Overlapping scores If the overlap ratio between the tooth mask and the adjacent tooth mask is high, the overlap score will be higher. high;

[0115] definition: .

[0116] Step S103, intraoral scanning 3D modeling and tooth position entity construction, includes:

[0117] Step S1031, Tooth / Gingival Segmentation and Single Tooth Mesh Extraction: A 3D segmentation network can output per-vertex label (tooth / gingiva / noise) and per-tooth instance. Alternatively, tooth crowns can be segmented based on curvature + normal variation + topographic watershed, and single-tooth instance meshes can be obtained by adjacency graph clustering. .

[0118] Step S1032, Tooth position FDI marking: Extract the centroid of each tooth crown. The sequence is obtained by sorting according to the center line of the dental arch; the sequence is mapped to FDI by the midline (incisor region) and quadrant division rules (missing teeth are skipped with placeholders); the tooth position entity data structure is output.

[0119] Step S1033, 3D key point extraction (for registration):

[0120] For each tooth Extraction: The highest point of the crown (maximum value along the normal to the occlusal plane) is used as the approximate point of the cusp / incisal edge. Crown boundary contour point set (The junction with the gingiva or the cervical margin).

[0121] In step S104, 2D-3D cross-modal registration is performed, and an approximate panoramic projection model is implemented. Panoramic imaging suffers from nonlinear scanning and distortion. This embodiment provides an engineering-feasible arch-parameter projection model, establishing a 3D-to-2D mapping with the dental arch length as the horizontal axis, and robustly optimizing the fitting parameters. Specifically, this includes:

[0122] Step S1041: Establish 3D dental arch parameter coordinates:

[0123] Take the centroids of all crowns Fit a three-dimensional spline curve as the centerline of the dental arch Define the arc length parameter for each tooth. (exist (projected arc length), and vertical coordinates (Relative occlusal plane height).

[0124] Step S1042, Panoramic Projection Model: Define the projection function : For 3D points First perform rigid body transformation Then mapped to panoramic pixel coordinates :

[0125] ,

[0126] ,

[0127] in For point Corresponding dental arch parameter coordinates; parameter set ,

[0128] This model provides nonlinear lateral scanning. With slight tilt Its fitting ability can cover most engineering errors in panoramic images.

[0129] Step S1043: Register and optimize the objective function, including robust terms and order constraints:

[0130] Take the mask for each tooth from the 2D panoramic film. The center of mass as a 2D key point In 3D, capture the key points of each tooth. ;

[0131] Key points: ;

[0132] in For Huber robust loss, threshold Pixel.

[0133] Contour item: 2D mask for each tooth Perform distance transformation to obtain 3D neck contour point set Sample several points : ;

[0134] Tooth position sequence constraint: Maintain monotonicity along the dental arch sequence. ,in The minimum interval is 2 pixels.

[0135] Overall goal: The weights can be set as follows: .

[0136] Step S1044, Solution Strategy and Stopping Conditions:

[0137] initialization: , Align with the arch plane (PCA aligns with the occlusal plane), then roughly align with the midline of the dental arch; Estimated using the ratio of dental arch length to image width and height. Center-align; Initially 0;

[0138] Iteration: Iteratively minimize using Levenberg-Marquardt or gradient descent ;

[0139] Termination: Number of iterations ≥ 50 or relative decrease .

[0140] Step S1045: Register confidence components Calculate the residual at each critical point of each tooth. (pixels). Define the overall residual mean. Residual difference per tooth The executable mapping is given as follows:

[0141] ,in Pixels Pixel.

[0142] when or At this time, the system's registration needs to be reviewed, and a conservative design strategy is triggered.

[0143] In step S105, image and registration credibility index synthesis and propagation (CRI) are performed.

[0144] In the process of automatically designing removable partial dentures, this invention considers both panoramic image quality and 2D–3D cross-modal registration reliability, incorporating image reliability components. Registration confidence components Synthesized into tooth-by-tooth images and registration reliability indicators This is used to characterize the overall reliability of radiographic bone loss (RBL) results at this tooth position in engineering and clinical applications.

[0145] Step S1051, Method for synthesizing the reliability index: For each tooth position The comprehensive credibility index is defined as follows: ;

[0146] in: This indicates the quantification of unreliability at the image level.

[0147] This represents the quantification result of unreliability at the cross-modal registration level;

[0148] ∈[0,1], the larger the value, the lower the reliability of the RBL result corresponding to that tooth position.

[0149] The above synthesis method uses a risk union model to ensure that if there is significant unreliability in any part of the image or registration process, the overall credibility of the tooth position is judged to be low.

[0150] Step S1052, Engineering semantics of credibility index:

[0151] In this invention, the image and registration reliability index It has the following engineering semantics:

[0152] 1) This is not a measurement error or confidence interval in the statistical sense;

[0153] 2) It is not used for smoothing, correcting, or regressing RBL values;

[0154] 3) This is used to characterize the level of risk involved when using the RBL results for automated denture design decisions.

[0155] when A lower value indicates good image quality and stable registration for that tooth position, and its RBL result can be used as a routine design constraint input; when A higher value indicates that the RBL result for that tooth position has high uncertainty and is not suitable for high-risk design decisions.

[0156] Step S1053, Credibility Propagation and Directional Risk Expression:

[0157] In an alternative implementation, the invention allows the confidence index to be further refined into direction-dependent confidence components to support more granular design constraints:

[0158] Near-Middle Confidence Ratio (CRI): Calculated from near-middle RBL measurement instability and corresponding mapping residuals;

[0159] Far-to-mid reliability (CRI): calculated from the far-to-mid RBL measurement instability and the corresponding mapping residual.

[0160] The directional confidence level can be used in subsequent designs to limit the use of undercuts or retainer arrangements in specific directions, thereby avoiding the application of excessive retaining forces in high-risk directions.

[0161] Step S1054, Credibility Threshold and Design Trigger Mechanism:

[0162] To facilitate engineering implementation, this invention sets at least one set of exemplary thresholds for the credibility index to trigger different design strategies:

[0163] when ,For example When the value is 0.3: the RBL result for this tooth position is considered reliable, and it can be used in abutment tooth selection and retention design according to conventional strategies; when ≤ < ， For example When the value is 0.6: the tooth position is considered to have a moderate confidence risk, and the system automatically adopts a conservative design strategy, including reducing the undercut depth, limiting the type of retention, or increasing the sharing of burden between the abutment and the denture base; when ≥ When: Mark the tooth position as requiring review. Its RBL results are not directly used for key design decisions, and the electronic design sheet will prompt for supplementary checks or manual confirmation.

[0164] The above thresholds can be adjusted according to clinical preferences or system configuration. This embodiment is for illustrative purposes only and is not limited to this. 0.3 and It is 0.6.

[0165] Step S1055, The role of credibility metrics in the design process:

[0166] Image and registration reliability index Its role in the system of this invention includes, but is not limited to:

[0167] As a risk penalty factor in the abutment tooth suitability score;

[0168] As a constraint factor for calculating the upper limit of the inverted concave depth;

[0169] As a criterion for restricting the selection and combination of fixation body types;

[0170] This serves as the basis for risk warnings and review markings in electronic design sheets.

[0171] By explicitly disseminating confidence information during the design process, this invention avoids outputting overconfident automated design results when image quality is insufficient or registration is unstable.

[0172] Step S106 involves constraint-driven RPD design based on radiographic bone loss and reliability indices.

[0173] After completing the tooth-by-tooth radiographic bone loss (RBL) assessment and the image and registration reliability index (CRI) calculation, RBL+CRI is used as the core constraint input to drive the selection of abutment teeth, control of undercut utilization, and configuration of retention for removable partial dentures (RPD), thereby balancing stability and clinical safety in the automated design process.

[0174] Step S1061, Abutment tooth suitability assessment:

[0175] opposing tooth position Define standardized variables: radiographic bone loss index ∈[0,1]; Image and registration reliability index ∈[0,1]; Looseness Normalization Inclination angle (degree), normalization Medical record risk markers The presence of caries / periapical region / severe restorations, etc. The value is 1.

[0176] Abutment tooth score: ,

[0177] Output category: Forbidden abutment tooth: RBL> and or and .

[0178] Requires review: If the image or registration is unreliable, the system can suggest additional examinations or manual confirmation.

[0179] Recommended abutment teeth: And it does not require review.

[0180] Optional abutment teeth: other cases.

[0181] When a tooth position is marked as requiring review, its corresponding RBL result is not directly used for high-risk design decisions, and a prompt is given in the electronic design sheet.

[0182] Step S1062, utilize the upper limit of the undercut depth:

[0183] For each candidate abutment tooth, define an upper limit for the undercut utilization depth. (Unit: mm):

[0184] ,

[0185] Unit: mm. Forced verification is required when necessary. (Conservative strategy).

[0186] Step S1063, Ring Type Restriction Rule Table:

[0187] Based on the combined states of RBL and CRI, explicit restrictions are imposed on the type and strength of the retainer, as shown in Table 1.

[0188] Table 1. Restriction Rules Table

[0189]

[0190] The specific name of the retainer is provided by the component library. In this embodiment, the retention strength and type are limited by RBL+CRI, rather than by a specific structural name.

[0191] Step S1064, Kennedy Classification and Structural Constraints:

[0192] The Kennedy class (I / II / III / IV) is automatically identified based on the position of the missing tooth. For the free end (Class I / II), the following rule is added: if the RBL or CRI of the main abutment tooth on the free end side is high, the coverage area of ​​the denture base is expanded first, and the stress relief retention strategy is used first.

[0193] Step S1065, Basement Coverage and Connector Selection, depends on intraoral scan soft tissue parameters:

[0194] Extracted from intraoral scan: alveolar crest line Boundary of the vestibular ditch , tethered activity area Denture base boundary formation: offset towards the buccal and lingual sides from the alveolar ridge crest line until approaching the vestibular sulcus boundary. Maintain a safety margin (e.g., 2mm) ahead; avoid the tethered area. ;Free end region according to Weighted Expanded Coverage: Coverage Factor: ,

[0195] in This represents the average radiographic bone loss index of the abutment teeth adjacent to the defect.

[0196] Connector selection rules:

[0197] If the thickness of the palatal mucosa and the area of ​​mobility allow, the form of connector that covers more stable areas should be preferred; if there is a highly mobile area, the avoidance form should be used instead.

[0198] The method disclosed in this embodiment also includes step S107: performing compliance verification and automatic repair on the automated design results based on manufacturing rules and clinical safety to obtain the electronic design sheet and manufacturable design data of the removable partial denture.

[0199] The compliance verification and automatic repair in step S107 are used to prevent the output of unmanufacturable or unsafe solutions.

[0200] Verification of the solution: Interference verification: Minimum gap between component and tooth / soft tissue ≥ 0.2mm; Minimum thickness: Connector / clasp arm thickness ≥ manufacturing threshold (e.g., ≥ 1.0mm or as per process configuration); Retention strength verification: Undercut depth must not exceed Conservative verification of credibility index: if the critical abutment tooth If the output is negative, the fixation strength needs to be reviewed and automatically downgraded.

[0201] The automatic repair strategy adopts the principle of minimal modification: first, reduce the depth of the indentation; then, replace the type of retainer with a more conservative type; then, expand the base cover or increase the support to share the load; if it is still not satisfactory, output the reason for failure and prompts that manual processing is required.

[0202] Output electronic design sheet and manufacturing data:

[0203] Output includes: Design sheet (PDF / HTML / editable form): 2D dental arch diagram, component annotations, text descriptions, risk warnings (including contraindications / reasons for review), key parameter table; Design data package (JSON or equivalent structure): component topology (connection relationships), component parameters for each tooth position, boundary curve control points, undercut utilization depth and direction, etc.; Optional export software interface format for subsequent CAD / manufacturing.

[0204] The data structure in this embodiment is as follows:

[0205] ToothRecord (Record per tooth position):

[0206] tooth_id(FDI);

[0207] presence (0 / 1);

[0208] mobility_grade (0-3);

[0209] tilt_angle_deg(0-90) / tilt_direction;

[0210] disease_flag (0 / 1);

[0211] RBL_mesial, RBL_distal, RBL(0-1);

[0212] CRI_img(0-1), CRI_reg(0-1), CRI(0-1);

[0213] score (0-100);

[0214] status (recommended / optional / prohibited / requires review);

[0215] d_max_mm (0.10-0.50).

[0216] ComponentRecord (component):

[0217] component_type(clasp / rest / connector / base / teeth);

[0218] anchor_tooth_id(FDI);

[0219] Orientation (proximal / distal / cheek / tongue);

[0220] curve_ctrl_pts (spline control point array);

[0221] undercut_depth_mm;

[0222] connectivity.

[0223] In a specific scenario, such as free end loss (Kennedy I):

[0224] Panoramic radiographs showed: Tooth position 36: RBL CRI →Recommended abutment tooth; Tooth position 35: RBL CRI →Contraindicated abutment teeth.

[0225] Design Engine:

[0226] For abutment tooth assembly, tooth position 35 and the tooth with better support on the opposite side are preferred;

[0227] Upper limit of undercut at tooth position 35 ;

[0228] The retention type should be selected as conservative / stress-relieving within the range of 0.45–0.65;

[0229] Free terminal baseplate coverage factor >1. Expand coverage and reduce the burden on the abutment teeth.

[0230] Output: The design sheet clearly states "Contraindications for tooth position 36: Insufficient bone support and high confidence level".

[0231] In another scenario, image overlap leads to a high credibility index:

[0232] A certain tooth position has low segmentation confidence and high registration residuals: .

[0233] The system outputs: "This tooth position needs to be reviewed", and forces the upper limit of the undercut to be ≤0.25mm. At the same time, it prompts for supplementary inspection or manual confirmation on the design sheet.

[0234] The technical effects of this implementation are as follows:

[0235] 1. Quantify the bone support information from panoramic radiographs tooth by tooth and map it to 3D tooth positions to reduce reliance on subjective recording;

[0236] 2. By employing uncertainty propagation and conservative strategies, the risk of misjudgment caused by image distortion / overlap can be reduced;

[0237] 3. The upper limit of fixation strength and type restrictions based on bone support constraints make the automatic design output more controllable;

[0238] 4. Output evidence-based design sheets to improve interpretability and clinical usability;

[0239] 5. It can reduce the risk of overloading abutment teeth and lower the probability of rework (which can be verified in clinical or technical processes through the "number of trial adjustments / rework rate" indicator).

[0240] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A method for designing removable partial dentures with cross-modal registration, characterized in that, include: Acquire two-dimensional panoramic radiographs and three-dimensional intraoral scans of the patient; The panoramic film is subjected to tooth-by-tooth detection and segmentation. Based on the segmentation results, the reference point of the cementoenamel junction, the apex point, and the mesial and distal ridge apex of each tooth are extracted. The radiographic bone loss ratio of each tooth is calculated based on the reference point, the apex point, and the ridge apex. Based on the segmentation results, the image confidence component of each tooth position in the panoramic film is obtained. The intraoral scan data is segmented into tooth structure and gingiva to construct a three-dimensional tooth position entity, and the three-dimensional key points and cervical margin contour point set of each tooth are extracted. A parametric projection model of the dental arch is established based on the three-dimensional key points and cervical margin contour point set of each tooth. Cross-modal registration is performed on the panoramic radiograph and intraoral scan data. The registration residual is obtained by solving the projection model parameters through optimization algorithm, and the registration confidence component of each tooth position is calculated based on the registration residual. The image confidence component of each tooth position in the panoramic film is fused with the registration confidence component to generate an image and registration confidence index. The proportion of radiographic bone loss and the reliability index of each tooth position are used as design constraints to automate the design of removable partial dentures.

2. The method for designing removable partial dentures with cross-modal registration according to claim 1, characterized in that, The calculation of the radiographic bone loss percentage for each tooth includes: Based on formula Calculate the root length of the tooth; Based on formula Calculate the near-center height; Based on formula Calculate the distance to mid-distance; The proportions of mesial and distal bone loss are calculated based on the root length, mesial height, and distal height of the tooth position. The larger of the mesial and distal bone loss proportions is selected as the radiographic bone loss proportion of the tooth; in, For root length, For the root tip, As a reference point for the cementoenamel junction, Near to medium altitude, At a distance of mid-altitude, To project onto the unit vector of the major axis, It is the proximal bony ridge. It is the distal ridge of the bone.

3. The method for designing removable partial dentures with cross-modal registration according to claim 1, characterized in that, The factors used to calculate the image confidence component in the panoramic film for each tooth position based on the segmentation results include: Confidence of tooth segmentation, clarity of tooth margin gradient, and overlap score of adjacent tooth masking; The formula for calculating the image credibility component is as follows: ,in, For image credibility components, The confidence level for tooth position segmentation. For the clarity of the gradient at the tooth margin, The overlap score of adjacent tooth position masks is calculated.

4. The method for designing removable partial dentures with cross-modal registration according to claim 1, characterized in that, The establishment of a parametric projection model of the dental arch based on the three-dimensional key points and cervical margin contour point set of each tooth includes: 3D points Using formula Perform a rigid body transformation, and then convert the transformation result... Mapped to panoramic pixel coordinates ; The mapping function is: ， ， in and For point The corresponding dental arch parameter coordinates, For arc length parameters, For height parameters, Let be a rotation matrix. It is a translation vector. This is the offset. The coefficient of the linear term, The coefficient of the quadratic term, As the vertical baseline, This is the vertical scaling factor. For coupling term coefficients, parameter set .

5. The method for designing removable partial dentures with cross-modal registration according to claim 4, characterized in that, The objective function for obtaining the registration residual by solving the projection model parameters through an optimization algorithm includes: ， in, , and All are weighting coefficients. For key point error, For contour error, As a constraint used to maintain the monotonicity of tooth position sequence; ， ， ， in, The total error of the keypoint matching items. This represents the total error of the contour matching item. This is the total error of the constraint term used to maintain the monotonicity of tooth position order. For Huber's robustness loss, For collection points, For the minimum interval, For the first The coordinates of the two-dimensional key points corresponding to each tooth on a two-dimensional panoramic image. For the parameterized projection function of the dental arch, For the first The coordinates of the three-dimensional key points corresponding to each tooth on the three-dimensional intraoral scan model. For the first The signed distance field function of a tooth on a two-dimensional panoramic film. For the first The coordinates of the three-dimensional key points corresponding to each tooth on the three-dimensional intraoral scan model. These are the horizontal coordinates of the two-dimensional point after projection.

6. The method for designing removable partial dentures with cross-modal registration according to claim 1, characterized in that, The formula for calculating the credibility index is as follows: ， in, As a credibility indicator, For image credibility components, To register the credibility component, The larger the value, the lower the reliability of the RBL result corresponding to that tooth position.

7. The method for designing removable partial dentures with cross-modal registration according to claim 1, characterized in that, The role of the image and registration reliability index includes: Risk penalty factor used in abutment tooth suitability scoring; Constraint factors used for calculating the upper limit of the depth for undercutting; Criteria used for the selection and combination restrictions of fixation body types; Used as the basis for risk warnings and review markings in electronic design sheets.

8. The method for designing removable partial dentures with cross-modal registration according to claim 7, characterized in that, The formula for calculating the suitability score of the abutment teeth is as follows: ， in, For the proportion of radiographic bone loss, As a credibility indicator, The normalized result for looseness, This is the normalized result for the tilt angle. This serves as a risk marker for medical records.

9. The method for designing removable partial dentures with cross-modal registration according to claim 8, characterized in that, The constraint factor calculated using the upper limit of depth for the indentation is: 。 10. The method for designing removable partial dentures with cross-modal registration according to claim 1, characterized in that, Following the step of automating the design of removable partial dentures using the proportion of radiographic bone loss and reliability indices for each tooth position as design constraints, the procedure further includes: The steps involve verifying and automatically repairing the automated design results based on manufacturing rules and clinical safety to obtain an electronic design sheet and manufacturable design data for removable partial dentures.

Images