AI-assistance-based rapid adaptation method for pre-crown formation of children

Abstract

The invention discloses an AI-assistance-based rapid adaptation method for pre-crowning of children, and belongs to the field of digital diagnosis and treatment of the dental department of children. The method comprises the following steps: 1, importing oral scanning data (. Stl / . Dcm / . Ply format) of a child patient; 2, a doctor selects target diseased teeth in the three-dimensional model; 3, performing AI segmentation on the dental crown, automatically positioning an adjacent point and a cheek-tongue highest point of a single diseased tooth, calculating a near-far intermediate diameter and a cheek-tongue diameter, and positioning a measurement point and calculating the size after AI simulation segmentation positions of a plurality of continuous diseased teeth; 4, simulation matching is conducted on the AI according to the size and the preformed crown model; 5, outputting a TOP3 model, a matching degree score and a three-dimensional superposition comparison graph; and 6, clinically verifying the adaptive effect. According to the method, manual errors are eliminated, second-level matching is achieved, multiple brands are adapted, the crown edge adaptability is improved, the matching time is shortened to 3 seconds per tooth, the measurement error is + / -0.05 mm, and the first-time adaptation rate is larger than or equal to 92%.

Description

Technical Field

[0001] This invention relates to the field of AI-assisted therapeutic engineering technology, specifically to an AI-assisted method for rapid fitting of pre-formed crowns in children. Background Technology

[0002] In pediatric dental care, the fitting of preformed crowns is a crucial step, and existing technical solutions mainly fall into two categories:

[0003] One method is the traditional method, where doctors manually measure plaster models or intraoral photographs, obtain tooth dimensions using calipers, and then manually match the model to a pre-formed crown size chart, or directly visually estimate tooth dimensions to select the pre-formed crown model.

[0004] Secondly, digital improvements have been made. Some clinics use software such as 3Shape to measure the size of a single tooth, but doctors still need to manually compare the measurements with a model database.

[0005] However, existing technologies have obvious drawbacks:

[0006] 1. Inefficient: Each tooth measurement, combined with manual comparison, takes 5-10 minutes, and the low cooperation of children will further prolong the operation time;

[0007] 2. The subjective error is large. The judgment of gingival margin morphology and the location of measurement points both depend on the doctor's experience, and different doctors are prone to differences in operation.

[0008] 3. Insufficient matching accuracy, failure to consider the three-dimensional geometric features of the teeth (such as buccal and lingual curvature), resulting in poor marginal fit of the preformed crown;

[0009] 4. It lacks dynamic adaptation capabilities and cannot automatically adjust the matching algorithm according to the design differences of different brands of pre-formed crowns (such as 3M, Stainless, etc.), thus limiting its applicability. Summary of the Invention

[0010] (a) Technical problems to be solved

[0011] To address the shortcomings of existing technologies, this invention provides an AI-assisted method for rapid fitting of pre-formed crowns for children, solving the technical problems mentioned in the background section.

[0012] (II) Technical Solution

[0013] To achieve the above objectives, the present invention provides the following technical solution: a rapid fitting method for preformed crowns for children based on AI assistance, comprising the following steps:

[0014] (1) Data input: Import the oral scan data of the child. The oral scan data supports .stl, .dcm and .ply formats.

[0015] (2) Target tooth selection: The doctor selects the target tooth that needs to be fitted with a preformed crown on the displayed 3D model of the tooth;

[0016] (3) AI Feature Extraction: The crown of the target tooth is segmented using a professional segmentation technology trained on 2000+ children’s deciduous teeth scan data. If it is a single tooth that needs to be fitted with a preformed crown and has natural teeth in front and behind, the mesiodistal junction and buccal-lingual highest point of the treated tooth and the adjacent natural tooth are automatically located, and the mesiodistal diameter (three-dimensional spatial distance between two adjacent points) and buccal-lingual diameter (normal distance between the highest points of the buccal and lingual surfaces) are calculated. If it is two or more consecutive teeth that need to be treated, the AI ​​automatically simulates the most suitable segmentation position between each tooth based on the large model data, and then automatically locates the mesiodistal junction and the highest point of the buccal and lingual surfaces of each treated tooth, and calculates the mesiodistal diameter (three-dimensional spatial distance between two adjacent points) and buccal-lingual diameter (normal distance between the highest points of the buccal and lingual surfaces).

[0017] (4) Intelligent matching: Based on the key dimension results calculated in step (3) and the pre-formed crown model database, simulation and matching are performed using AI;

[0018] (5) Results output: Output the top 3 recommended pre-formed crown models, the matching score of each model, and display the three-dimensional overlay comparison chart;

[0019] (6) Clinical validation: The recommended preformed crowns were tested to verify the fit effect, such as the crown edge tightness.

[0020] Preferably, the training data used for segmenting the crown in step (3) is 2000+ children's deciduous teeth scan data to ensure segmentation accuracy.

[0021] Preferably, the pre-formed crown model database in step (4) contains pre-formed crown model information from at least two different brands, such as 3M and Stainless.

[0022] Preferably, the matching score in step (5) is calculated based on the geometric similarity algorithm between the preformed crown and the target tooth.

[0023] Preferably, when calculating the mesial and distal diameters and the buccal and tongue diameters in step (3), a three-dimensional spatial distance measurement method is used to ensure the accuracy of the dimensional data.

[0024] Preferably, the core indicator for clinical verification in step (6) is the coronal margin fit, and the fit must reach more than 90%.

[0025] Preferably, the oral scan data imported in step (1) needs to be preprocessed to remove noise data and ensure data quality.

[0026] Preferably, in step (3), when the AI ​​simulates the segmentation position of multiple consecutive affected teeth, the arrangement pattern and morphological characteristics of children's teeth in the large model data are referenced.

[0027] Preferably, the three-dimensional superimposed comparison image shown in step (5) can be enlarged and rotated, so that doctors can intuitively observe the fit between the preformed crown and the target tooth.

[0028] (III) Beneficial Effects

[0029] Compared with existing technologies, this invention provides an AI-assisted rapid fitting method for preformed crowns in children, which has the following beneficial effects:

[0030] 1. Significantly improves efficiency, reducing the matching time of a single tooth from 5-10 minutes in the traditional method to 3 seconds per tooth, significantly reducing treatment time and reducing the difficulty for children to cooperate;

[0031] 2. Eliminate human error by automatically locating measurement points and calculating dimensions using AI, reducing measurement error from the traditional ±0.3mm to ±0.05mm and improving data accuracy;

[0032] 3. Improved fitting accuracy, increasing the initial fitting rate from the traditional 60-70% to ≥92%, and fully considering the three-dimensional geometric features of teeth, effectively improving the crown margin fit, with a crown margin fit of up to 96% in clinical trials;

[0033] 4. Enhance brand compatibility and break through the limitations of traditional methods that only adapt to a single brand. Through brand rule base and dynamic adaptation model, it can automatically adapt to the design standards of different brands of preformed crowns to meet diverse clinical needs.

[0034] 5. The operation is convenient and intuitive. The output results include the top 3 models, matching score and three-dimensional overlay comparison chart, which makes it easy for doctors to make quick judgments and selections, reduces the reliance on doctors' experience and facilitates the promotion and application of the technology. Detailed Implementation

[0035] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] A rapid fitting method for preformed crowns for children based on AI assistance includes the following steps: (1) Data input: Import the oral scan data of the child, which supports .stl, .dcm and .ply formats; (2) Target tooth selection: The doctor clicks on the displayed three-dimensional model of the tooth to select the target tooth that needs to be fitted with a preformed crown; (3) AI feature extraction: The crown of the target tooth is segmented using a professional segmentation technology trained based on 2000+ children's deciduous tooth scan data; If it is a single tooth that needs to wear a preformed crown and has natural teeth in front and behind, the mesiodistal contact point and buccal-lingual highest point of the treated tooth and the adjacent natural tooth are automatically located, and the mesiodistal diameter (the three-dimensional spatial distance between the two adjacent points) is calculated. (3) Intelligent matching: Based on the key dimension results calculated in step (3) and the preformed crown model database, the AI ​​simulates and matches the results; (4) Results output: Output the recommended TOP3 preformed crown models, the matching degree score corresponding to each model, and display the three-dimensional superimposed comparison map; (5) Clinical Verification: The recommended preformed crowns were tested to verify the fit, such as the crown margin fit. In step (3), the training data used for dividing the crowns consisted of 2000+ children's deciduous teeth scans to ensure the accuracy of the division. In step (4), the preformed crown model database contained information on at least two different brands of preformed crowns, such as 3M and Stainless. In step (5), the matching score was calculated based on the geometric similarity algorithm between the preformed crown and the target tooth. In step (3), the mesiodistal diameter and buccal-lingual diameter were calculated using a three-dimensional spatial distance measurement method to ensure the accuracy of the dimensional data. In step (6), the core indicator for clinical verification was the crown margin fit, and the fit had to reach 9. The oral scan data imported in step (1) needs to be preprocessed to remove noise data and ensure data quality. In step (3), when the AI ​​simulates the segmentation position of multiple consecutive teeth, the arrangement rules and morphological characteristics of children's teeth in the large model data are referenced. The three-dimensional superimposed comparison image displayed in step (5) can be enlarged and rotated, which makes it easier for doctors to intuitively observe the fit between the preformed crown and the target tooth. This invention greatly improves efficiency, shortens the matching time of 5-10 minutes per tooth in the traditional method to 3 seconds / tooth, significantly reduces the treatment time, and reduces the difficulty of children's cooperation. It eliminates human error and reduces the measurement error from the traditional ±0.3mm to ±0mm by automatically locating the measurement point and calculating the size through AI.The 0.05mm reduction improves data accuracy and fitting precision, increasing the initial fitting rate from the traditional 60-70% to ≥92%. It also fully considers the three-dimensional geometric characteristics of teeth, effectively improving crown margin fit; clinical trials have shown a crown margin fit of up to 96%. Furthermore, it enhances brand compatibility, overcoming the limitations of traditional methods that only fit a single brand. Through a brand rule base and dynamic fitting model, it automatically adapts to different brands of pre-formed crown design standards, meeting diverse clinical needs. The operation is convenient and intuitive, with output including the top 3 models, matching score, and a 3D overlay comparison image, facilitating quick judgment and selection by doctors, reducing reliance on doctor experience, and promoting the widespread application of the technology.

[0037] Example:

[0038] An Example of an AI-Assisted Rapid Fitting Method for Preformed Crowns in Children

[0039] This embodiment takes the treatment scenario of a child's left lower first deciduous molar (tooth position marked as lower right E) requiring a preformed crown as an example to explain in detail the specific implementation process of the present invention, ensuring that those skilled in the art can clearly understand and reproduce the method.

[0040] I. Data Acquisition and Import

[0041] Scanning procedure: The child's oral cavity was scanned using a 3Shape intraoral scanner, with a focus on collecting complete data of the mandibular dental arch, including the left lower first deciduous molar. During the scanning process, the probe was kept in close contact with the tooth surface to avoid data loss or blurring due to slight movement of the child. After the scan was completed, a standard three-dimensional model of the teeth was generated using the scanner's accompanying software.

[0042] Data export and import: In 3Shape software, export the generated 3D tooth model in .stl format (or .dcm or .ply format according to actual needs); open the AI-assisted adaptation system corresponding to this invention, select the exported .stl file through the system's "Data Import" function, and complete the import of the child's oral scan data. The system will automatically load and display the 3D tooth model.

[0043] II. Target Tooth Selection

[0044] In the operation interface of the AI-assisted adaptation system, the doctor clicks on the lower left first deciduous molar on the 3D model with the mouse (the system interface will mark the tooth position, with the lower left first deciduous molar being marked as "lower right E"). After clicking, the system will automatically highlight the tooth. After confirming the selection, a prompt box will pop up saying "Target tooth has been selected. Do you want to enter the AI ​​processing flow?" The doctor clicks "confirm" to proceed to the next step.

[0045] III. AI Feature Extraction and Size Calculation

[0046] Crown segmentation: The system activates an AI tooth segmentation network, trained on scan data of over 2000 children's deciduous teeth, which can accurately identify and segment the crown portion of the target tooth. After segmentation, the system automatically hides interfering areas such as the gums and other teeth outside the crown, displaying only the 3D model of the crown of the lower left first deciduous molar, facilitating subsequent feature point localization.

[0047] Determination of affected tooth condition and localization of feature points: The system uses AI algorithms to analyze the adjacent teeth of the target affected tooth, determining that the tooth is a single tooth requiring a preformed crown and has natural teeth in both front and back (mesial: left lower second primary premolar; distal: left lower second primary molar). Then, it automatically locates and measures the affected tooth.

[0048] Mesial point: Locate the highest point on the distal surface of the mesial adjacent tooth (left mandibular second premolar), marked with a red dot;

[0049] Distal point: Locate the highest point on the mesial surface of the distal adjacent tooth (left mandibular second deciduous molar), marked with a red dot;

[0050] Buccal protrusion: The highest point of the outline on the buccal surface of the target tooth, marked with a red dot;

[0051] Lingual protrusion: The highest point of the lingual surface of the target tooth, marked with a red dot.

[0052] Critical dimension calculation: The system calculates critical dimensions based on the located marker points using a three-dimensional spatial distance measurement algorithm.

[0053] Mesial-distal diameter: The three-dimensional straight-line distance between the mesial and distal midpoints was calculated, and the result was 5.2 mm;

[0054] Buctolingual diameter: The normal distance between the buccal and lingual protrusions (perpendicular to the occlusal surface of the teeth) was calculated, and the result was 4.8 mm;

[0055] Once the calculation is complete, the dimensional data will be automatically displayed in the "Dimensional Parameters" panel of the system interface.

[0056] IV. Pre-formed crown model matching

[0057] Brand Selection: The system interface provides a list of preformed crown brands, including mainstream brands such as 3M and Stainless. Doctors can click to select "3M Brand Library" based on commonly used brands in clinical practice and the needs of the child's treatment. The system will automatically load the model database of 3M preformed crowns for children's deciduous molars (including parameters such as the standard range of mesiodistal diameter and buccal-lingual diameter for each model).

[0058] AI Intelligent Matching: The system activates the intelligent matching engine, combines the calculated key dimensions of the target tooth (mesiodistal diameter 5.2mm, buccal-lingual diameter 4.8mm), calls the geometric similarity algorithm to compare the three-dimensional morphological features of the tooth with the three-dimensional models of various preformed crowns in the 3M brand library, and at the same time refers to the design features of 3M preformed crowns in the brand rule library (such as the buccal-lingual curvature of the crown body, gingival margin morphology, etc.), and adjusts the matching weight through dynamic adaptation model to complete the model selection.

[0059] V. Matching Result Output

[0060] After the system completes the matching, the following content will be output in the "Results Display" area of ​​the operation interface:

[0061] Top 3 Recommended Models: Sorted from highest to lowest matching degree, the recommended models under the 3M brand are "3M-SSC-4" (matching degree 98%), "3M-SSC-5" (matching degree 92%), and "3M-SSC-3" (matching degree 89%).

[0062] 3D overlay comparison image: For each recommended model, a 3D overlay model image of the preformed crown and the target tooth crown is generated. Different colors are used to distinguish the tooth (white) and the preformed crown (blue) in the image. Doctors can drag the mouse to zoom in and rotate the overlay image to intuitively observe the fit between the preformed crown and the tooth in the buccal and lingual and mesiodistal directions.

[0063] Matching details: Click on each recommended model to view the deviation value between the model and the key dimensions of the affected tooth (e.g., "3M-SSC-4" mesiodistal diameter deviation +0.1mm, buccal-lingual diameter deviation -0.05mm) and geometric similarity analysis report.

[0064] VI. Clinical Validation and Adaptation Adjustment

[0065] Preformed crown trial fitting: Based on the system's recommendations, the doctor will prioritize the "3M-SSC-4" preformed crown model with the highest matching degree and perform a trial fitting on the child's left lower first deciduous molar to check the placement, marginal fit, and occlusal contact of the preformed crown.

[0066] Fit test: The gap between the pre-formed crown margin and the cervical margin of the affected tooth was gently probed with a dental probe, and intraoral photography was taken to assist in the observation. The test results showed that the crown margin fit reached 96%, which met the clinical fitting standard.

[0067] Confirmed Fit: The child experienced no significant discomfort during the trial fitting, and the occlusal function was normal. The doctor confirmed that the "3M-SSC-4" model preformed crown was successfully fitted, completing this rapid fitting procedure for the child's preformed crown.

[0068] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A rapid fitting method for preformed crowns for children based on AI assistance, characterized in that, Includes the following steps: (1) Data input: Import the oral scan data of the child. The oral scan data supports .stl, .dcm and .ply formats. (2) Target tooth selection: The doctor selects the target tooth that needs to be fitted with a preformed crown on the displayed 3D model of the tooth; (3) AI Feature Extraction: The crown of the target tooth is segmented using a professional segmentation technology trained on 2000+ children’s deciduous teeth scan data. If it is a single tooth that needs to be fitted with a preformed crown and has natural teeth in front and behind, the mesiodistal junction and buccal-lingual highest point of the treated tooth and the adjacent natural tooth are automatically located, and the mesiodistal diameter (three-dimensional spatial distance between two adjacent points) and buccal-lingual diameter (normal distance between the highest points of the buccal and lingual surfaces) are calculated. If it is two or more consecutive teeth that need to be treated, the AI ​​automatically simulates the most suitable segmentation position between each tooth based on the large model data, and then automatically locates the mesiodistal junction and the highest point of the buccal and lingual surfaces of each treated tooth, and calculates the mesiodistal diameter (three-dimensional spatial distance between two adjacent points) and buccal-lingual diameter (normal distance between the highest points of the buccal and lingual surfaces). (4) Intelligent matching: Based on the key dimension results calculated in step (3) and the pre-formed crown model database, simulation and matching are performed using AI; (5) Results output: Output the top 3 recommended pre-formed crown models, the matching score of each model, and display the three-dimensional overlay comparison chart; (6) Clinical validation: The recommended preformed crowns were tested to verify the fit effect, such as the crown edge tightness.

2. The method for rapid fitting of pre-formed crowns for children based on AI assistance according to claim 1, characterized in that, The training data used for segmenting the crowns in step (3) consisted of 2000+ children's deciduous teeth scan data to ensure segmentation accuracy.

3. The method for rapid fitting of pre-formed crowns for children based on AI assistance according to claim 1, characterized in that, The pre-formed crown model database in step (4) contains pre-formed crown model information from at least two different brands, such as 3M and Stainless.

4. The method for rapid fitting of pre-formed crowns for children based on AI assistance according to claim 1, characterized in that, The matching score in step (5) is calculated based on the geometric similarity algorithm between the preformed crown and the target tooth.

5. The method for rapid fitting of pre-formed crowns for children based on AI assistance according to claim 1, characterized in that, In step (3), when calculating the mesial and distal diameters and the buccal and lingual diameters, a three-dimensional spatial distance measurement method is used to ensure the accuracy of the dimensional data.

6. The method for rapid fitting of pre-formed crowns for children based on AI assistance according to claim 1, characterized in that, The core indicator for clinical validation in step (6) is the coronal margin fit, and the fit must reach more than 90%.

7. The method for rapid fitting of pre-formed crowns for children based on AI assistance according to claim 1, characterized in that, The oral scan data imported in step (1) needs to be preprocessed to remove noise data and ensure data quality.

8. The method for rapid fitting of pre-formed crowns for children based on AI assistance according to claim 1, characterized in that, In step (3), when the AI ​​simulates the segmentation position of multiple consecutive affected teeth, the arrangement pattern and morphological characteristics of children's teeth in the large model data are referenced.

9. The method for rapid fitting of pre-formed crowns for children based on AI assistance according to claim 1, characterized in that, The three-dimensional overlay comparison image shown in step (5) can be enlarged and rotated, making it easier for doctors to visually observe the fit between the preformed crown and the target tooth.