Artificial intelligence oral operation consumable usage evaluation optimization method

Abstract

The invention discloses an artificial intelligence oral surgery consumable use evaluation optimization method, which comprises the following steps: acquiring three-dimensional image data of an oral cavity of a patient, including adjacent tooth marrow cavity gray texture and gray jump characteristics of a pentaplet region; performing denoising processing and edge enhancement on the three-dimensional image data to obtain a target CBCT image; based on an adaptive particle swarm algorithm, five to-be-selected planting site images in the pentaplet region of the target CBCT image are extracted, and a current to-be-selected planting site is calculated and determined as a target planting site; then selecting three to five target planting sites from the pentaad area and a boundary area image of the pentaad area where the current target planting site is located; and determining a final target combined implant conjunct according to the bearing evaluation index of the pentaplet integral implant. According to the method, the planting scheme selection quality of the pentaplet integral implant in the operation is effectively improved, and a solid foundation is laid for subsequent planting material selection.

Description

Technical Field

[0001] The present invention relates to the field of digital implant consumables resource allocation, and in particular to an artificial intelligence-based oral surgery consumables usage evaluation and optimization method. Background Art

[0002] Dental implant technology is a surgical procedure that implants an artificial implant (usually a threaded cylindrical structure made of titanium alloy) into the alveolar bone. After it forms a stable osseointegration with the bone tissue, an abutment and crown are installed on top of the implant to replace missing teeth.

[0003] Implants, such as Straumann implants and Nobel Biocare implants, are core consumables used in oral implant surgery. They replace missing teeth and restore chewing function and aesthetics. Implant abutments, also known as implant abutments, are crucial components connecting the implant and the crown. There are various types, including adhesive abutments and locator abutments. Choosing the appropriate abutment for the restoration is tailored to the patient's specific needs.

[0004] During preoperative preparation, the doctor will conduct a comprehensive oral examination of the patient, including taking panoramic oral radiographs and cone beam CT scans, to understand the density, height, and width of the alveolar bone, as well as the condition of adjacent teeth and tissues, and to develop a personalized implant plan. However, studies have found that the above-mentioned personalized implant plans are often based on the doctor's experience, reliability, and safe implant conditions based on panoramic oral radiographs, cone beam CT scans, and other oral observation data. The doctor then selects the implant plan based on reliability and safe implant conditions, including the type of implant and the consumption of the implant and abutment.

[0005] However, it is clear that after years of development in digital reconstruction of dental implants, it is now possible to use software to evaluate oral implant conditions and develop corresponding implant plans. The ultimate goal of digital reconstruction of dental implants is to use digital equipment and technologies, such as cone-beam CT (CBCT), intraoral scanners, computer-aided design / computer-aided manufacturing (CAD / CAM) systems, to accurately collect, analyze and simulate patient oral conditions, develop personalized implant plans, and guide implant surgery and restoration production.

[0006] The digital reconstruction in existing technologies is based on the limitations of image detection methods and cannot fully achieve accurate judgment and analysis. In particular, the material selection of the pentad (three teeth from five adjacent lost teeth sites that meet the alveolar bone height requirements and the surrounding teeth are healthy) is crucial.

[0007] Further research and development revealed that it is impossible to evaluate and obtain a reliable implant plan by simply relying on doctors to observe the alveolar bone height of a single tooth without considering factors such as the reliability of the entire pentagonal implant and the overall alveolar absorption. Summary of the Invention

[0008] The purpose of the present invention is to provide an artificial intelligence-based oral surgery consumables usage evaluation and optimization method to solve the above-mentioned technical problems pointed out in the prior art.

[0009] The present invention provides an artificial intelligence-based oral surgery consumables usage evaluation and optimization method, comprising the following steps:

[0010] Acquire the patient's oral 3D imaging data, including the grayscale texture of the adjacent tooth medullary cavity and the grayscale jump characteristics of the pentad area;

[0011] Denoising and edge enhancement are performed on the three-dimensional image data to obtain a target CBCT image; based on an adaptive particle swarm algorithm, five images of candidate implant sites in the quintuplicate region of the target CBCT image are extracted; then, using a peak shape index of the target implant site, the compatibility between the single alveolar bone and the implant corresponding to the current candidate implant site is quantified to be greater than a standard compatibility threshold; if so, the current candidate implant site is determined as the target implant site; then, three to five target implant sites and boundary region images of the quintuplicate region where the current target implant site is located are selected from the quintuplicate region;

[0012] Perform pixel point analysis based on the target implant site image in the quintuplet area and the boundary area image of the quintuplet area, and calculate the quintuplet overall implant load evaluation index based on the grayscale texture of the adjacent tooth medullary cavity of the target implant site and the grayscale jump characteristics of the quintuplet area; then determine whether the current target combination implant is one or more; and determine the final target combination implant based on the quintuplet overall implant load evaluation index;

[0013] Combined implants are combined to form a quintuple area based on the final goal, and the implant length of the target implant site is calculated. Then, the initial abutment type is determined based on the overall implant load evaluation index of the quintuple area. Based on the selected abutment type and implant length, the material usage of the implant length and the abutment type are determined, and a three-dimensional visual implant plan is generated.

[0014] Preferably, as an implementable method, the adaptive particle swarm algorithm is used to extract five candidate implant site images in the quintuple region of the target CBCT image, and then the peak shape index of the target implant site is used to quantify whether the compatibility between the single alveolar bone and the implant of the current candidate implant site is greater than the standard compatibility threshold. If so, the current candidate implant site is determined as the target implant site, specifically including:

[0015] First, particle swarm initialization and adaptive parameter setting are performed: an initialization particle swarm is defined on the preprocessed quintuplet region image; each particle is defined to represent the position of a candidate implantation site and its corresponding adaptive parameters; the adaptive parameters include the position, velocity, and state parameters of the initialization particle;

[0016] During initialization, a local fitness function is defined for each particle, which is calculated based on the normalized value of the average grayscale value in the ROI area, the normalized value of the gradient amplitude in the ROI area, the normalized value of the texture feature, and the normalized value of the edge feature;

[0017] Perform peak shape index and fitness quantitative evaluation on the determined ROI area to calculate the fitness Ai of the current candidate site; summarize and calculate the fitness Ai of each candidate site, and judge its relationship with the fitness threshold of the Tadapt standard to screen candidate sites;

[0018] For each candidate site, compare its suitability Ai to see if it meets the standard suitability threshold of Ai ≥ Tadapt; retain the candidate sites that meet the conditions, and eliminate those that do not; record the number of qualified candidate sites: if the number of candidates is exactly 3, directly proceed to the subsequent evaluation step of the pentad integral implant load-bearing evaluation index; if the number of candidates is greater than 3, proceed to the evaluation step of the pentad integral implant load-bearing evaluation index of the multiple schemes of the candidate combination.

[0019] Finally, the target candidate ROI region of the target implantation site extracted from the quintuplet region and its corresponding boundary region image are determined.

[0020] Preferably, as an implementable method, the step of performing a quantitative evaluation of the peak shape index and the suitability of the determined ROI region to obtain the suitability Ai of the current candidate site specifically includes the following steps:

[0021] For each candidate implant site, obtain the grayscale value of each pixel in the jth slice of the ROI area ; Get the lowest grayscale value in the candidate ROI area; Determine the peak value of the candidate ROI area As a reference indicator;

[0022] Perform fitness calculation processing operations: ; in: : No. The fitness of each candidate site; : No. The planting site is in Layer slice The grayscale value of each pixel; : The lowest gray value in the candidate ROI area; : The average gray value of all pixels along the alveolar ridge top in the candidate ROI area; : is the weight coefficient; : A very small positive value to prevent division by zero.

[0023] Preferably, as an implementable embodiment; performing pixel point analysis based on the target implant site image in the quintuplet area and the boundary area image of the quintuplet area, and calculating the quintuplet overall implant load evaluation index in combination with the grayscale texture of the adjacent tooth medullary cavity of the target implant site and the grayscale jump characteristics of the quintuplet area; then judging whether the current target combined implant conjoint is one or more; determining the final target combined implant conjoint based on the quintuplet overall implant load evaluation index, specifically including:

[0024] Pixel feature extraction is performed on the ROI image of the target implant site and the corresponding boundary area image. When extracting features from the ROI image of the target implant site, the average grayscale value of the pixel features within the ROI is extracted. When extracting boundary features from the ROI image of the target implant site, the edge sharpness and grayscale uniformity of the adjacent tooth medullary cavity texture are extracted based on the boundary grayscale jump information.

[0025] The evaluation index of the pentad implant load was calculated based on the average gray value of the pixel features within the ROI, the edge sharpness, and the gray uniformity of the adjacent tooth medullary cavity texture. ;

[0026] in: It is an evaluation index for the overall implant load; is the average gray value of the target ROI area corresponding to any three target implantation sites; is the average value of edge sharpness in the boundary area corresponding to each of the three target planting sites; Grayscale uniformity value of the adjacent tooth medullary cavity texture corresponding to each of the three target implant sites; is the corresponding weight coefficient;

[0027] Based on the calculated pentad implant load evaluation index, candidate target combinations are screened and the final implant combination plan is determined. The specific operations are as follows:

[0028] First, a threshold comparison is performed, that is, the evaluation index C corresponding to each candidate target combination is compared with the minimum overall load threshold Tcarry: if the evaluation index C corresponding to the candidate target combination ≥ Tcarry, then the target combination is considered to meet the implant load requirements;

[0029] Then, it is determined whether the current target combined implant combination is one or more than one; and the final target combined implant combination is determined based on the five-unit overall implant load evaluation index;

[0030] If the current target implant combination is one, determine whether the overall implant load evaluation index of the quintuple is greater than the minimum overall load threshold. If so, determine that the current target implant combination is the final target implant combination;

[0031] If there are multiple target combination implants at present, determine whether the five-body overall implant load evaluation index corresponding to each target combination implant is greater than the minimum overall load threshold. If all are greater, establish a numerical arrangement list of load evaluation indexes and select the first screening combination processing condition to determine the final target combination implant or determine the final target combination implant according to the second screening combination processing condition.

[0032] Preferably, as an implementable embodiment, the first screening combination processing condition is to select the target combination implantation conjoint with the largest value in the current load evaluation index numerical arrangement list as the final target combination implantation conjoint.

[0033] Preferably, as an implementable embodiment; the second screening assembly processing condition is to set three priority screening levels, including a first priority screening level, a second priority screening level and a third priority screening level;

[0034] Among them, the first priority screening level is to execute according to the planting site "1-3-5" as the first preferred combination; the second priority screening level is to execute according to the planting site "1-X-5" as the second preferred combination; the third priority screening level is to execute according to the planting site "2-3-4" as the third preferred combination.

[0035] Preferably, as an implementable method, after "extracting five candidate implant site images in the quintuple region of the target CBCT image based on an adaptive particle swarm algorithm, and then quantifying whether the compatibility of the single alveolar bone and the implant corresponding to the current candidate implant site meets or exceeds the standard compatibility threshold value through the peak shape index of the target implant site, and if so, determining the current candidate implant site as the target implant site", it also includes performing an optimization processing operation:

[0036] The newly added first parameter includes the periodontitis inflammatory absorption coefficient P; the newly added second parameter includes the minimum implant spacing limit value D1;

[0037] The first parameter and the second parameter are introduced based on the iterative search and update operation of the adaptive particle swarm; an optimized local fitness function is defined for each particle, taking into account the local gray value, gradient, texture information and edge features; the optimized local fitness function is defined as: Fb=[F×(1− '×P)]× ;

[0038] ' is the predetermined adjustment coefficient; P is the periodontitis inflammation absorption coefficient;

[0039] is the spacing limit penalty factor, defined as follows:

[0040] If the distances between the current candidate site and its adjacent candidate sites are The minimum distance between implants is limited to D1. ;

[0041] If the distance between any adjacent candidate sites is The minimum distance between implants is limited to D1. .

[0042] Preferably, as an implementation plan, after determining the final target combination implantation conjoint according to the second screening combination processing condition, the method further includes: further detecting the implantation sites in the current final target combination implantation conjoint;

[0043] Determine whether the adjacent teeth at the current implant site are tilted;

[0044] If it is determined that the adjacent teeth at the current implant site are tilted, it is determined that the current final target combined implant joint encroaches on the implant space;

[0045] Filter out the current final target combination implant for further orthodontic or grinding treatment.

[0046] Accordingly, the present invention provides a storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-mentioned artificial intelligence oral surgery consumables usage evaluation and optimization method.

[0047] Compared with the prior art, the embodiments of the present invention have at least the following technical advantages:

[0048] Analysis of the artificial intelligence oral surgery consumables usage evaluation and optimization method provided by the present invention shows that in specific applications, first, three-dimensional image data of the patient's oral cavity is obtained, including the grayscale texture of the adjacent tooth medullary cavity and the grayscale jump characteristics of the pentagonal area; the three-dimensional image data is denoised and edge enhanced to obtain the target CBCT image;

[0049] Extract five candidate implant site images in the quintuplicate region of the target CBCT image based on an adaptive particle swarm algorithm, then quantify whether the compatibility of the single alveolar bone and the implant corresponding to the current candidate implant site meets or exceeds a standard compatibility threshold using the peak shape index of the target implant site. If so, determine the current candidate implant site as the target implant site, and then select three to five target implant sites from the quintuplicate region and the boundary region image of the quintuplicate region where the current target implant site is located;

[0050] Perform pixel point analysis based on the target implant site image in the quintuplet area and the boundary area image of the quintuplet area, and calculate the quintuplet overall implant load evaluation index based on the grayscale texture of the adjacent tooth medullary cavity of the target implant site and the grayscale jump characteristics of the quintuplet area; then determine whether the current target combination implant is one or more; and determine the final target combination implant based on the quintuplet overall implant load evaluation index;

[0051] Combine the final target implant combination to form a pentad area, and calculate the implant length of the target implant site;

[0052] Then, the initial abutment type is determined based on the overall implant load evaluation index of the pentagonal area; then, based on the selected abutment type and implant length, the material usage and abutment type of the implant length are determined, and a three-dimensional visual implant plan is generated.

[0053] The above technical solution utilizes an adaptive particle swarm algorithm; it also extracts five candidate implant site images in the quintuple region of the target CBCT image based on the adaptive particle swarm algorithm, and then quantifies the current candidate implant site, i.e., the selection of a single implant position, using the peak shape index of the target implant site;

[0054] Then, the overall implant stability is determined by factors such as the grayscale texture of the adjacent tooth medullary cavity and the grayscale jump characteristics of the pentad area. However, the stability of this overall implant site is different from the stability of a single implant, so studying the evaluation index of the pentad overall implant load is an important technical means. The above technical solution combines the adaptive particle swarm algorithm to realize the screening of implant sites in the pentad area. Compared with traditional manual marking, the algorithm dynamically adjusts the particle search strategy, taking into account the grayscale texture of the adjacent tooth medullary cavity and the difference in bone density, while locating the candidate sites efficiently and avoiding the local optimal trap. The introduction of the peak shape index quantification model further transforms the evaluation of the suitability of a single implant from qualitative judgment to quantitative analysis, which reduces the error rate of threshold judgment.

[0055] The above solution avoids the limitations of traditional single-implant stability assessment and utilizes the overall load-bearing evaluation index of the pentad. By integrating three-dimensional parameters: alveolar bone density (biomechanical basis), grayscale texture of the adjacent tooth medullary cavity (mechanical conduction balance), and grayscale jump characteristics, a holistic evaluation model is established. This model improves the accuracy of long-term stability assessments of the conjoined structure and effectively reduces the risk of bone resorption and mechanical complications. Through innovations in algorithm-driven, data fusion, and mechanical modeling, this technical solution achieves full-chain optimization from image analysis to surgical execution, improving implant success rates while establishing precise selection of implant consumables. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0057] Figure 1 This is an overall flow chart of an artificial intelligence method for evaluating and optimizing the use of oral surgery consumables provided by the present invention;

[0058] Figure 2 The original CBCT image collected by the present invention;

[0059] Figure 3 A diagram showing the process of particle analysis of the original CBCT image acquired by the present invention;

[0060] Figure 4 This is a practical schematic diagram of an implant combination in the CBCT image provided by the present invention;

[0061] Figure 5A schematic diagram of one combination of target combination implants in an artificial intelligence oral surgery consumables use evaluation and optimization method provided by the present invention;

[0062] Figure 6 A schematic diagram of another combination of target combination implants in an artificial intelligence oral surgery consumables use evaluation and optimization method provided by the present invention;

[0063] Figure 7 A schematic diagram of another combination selection of a target combination implant conjoined body in an artificial intelligence oral surgery consumables use evaluation and optimization method provided by the present invention;

[0064] Figure 8 A storage medium provided in Embodiment 5 of the present invention;

[0065] Label: processor 1110 ; communication interface 1120 ; memory 1130 ; computer storage medium 1140 . DETAILED DESCRIPTION

[0066] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0067] The present invention will be further described in detail below through specific embodiments in conjunction with the accompanying drawings.

[0068] Example 1 like Figure 1 As shown, the present invention proposes an artificial intelligence oral surgery consumables usage evaluation optimization method, including the following steps:

[0069] S10: Acquire three-dimensional oral imaging data of the patient, including grayscale texture of the adjacent tooth medullary cavity (i.e., stress on the alveolar bone connection between adjacent teeth, which reflects the health of the adjacent teeth) and grayscale transition characteristics of the pentad region (five adjacent teeth). Through the acquisition and subsequent analysis of the grayscale texture of the adjacent tooth medullary cavity and the grayscale transition characteristics of the pentad region, an overall implant bearing evaluation index for the pentad is constructed. For details, see step S30.

[0070] Specifically, the system (i.e., the AI-powered oral surgery medical consumables resource utilization assessment and optimization system described in Example 2) acquires CBCT images. Specifically, high-resolution images of the patient's oral region are acquired using CBCT (cone-beam CT) three-dimensional scanning technology. The system then transmits the data to obtain the CBCT images. The CBCT image data includes alveolar bone density distribution (represented by CT values ​​or bone density indices), grayscale texture of the adjacent tooth medullary cavity (i.e., stress on the alveolar bone connection between adjacent teeth), and even the health status of adjacent teeth (e.g., caries and wear), as well as grayscale transition characteristics within the pentad region.

[0071] S20: Denoising (e.g., Gaussian filtering) and edge enhancement (e.g., Laplacian operator) are performed on the three-dimensional image data to obtain a target CBCT image; then, based on the adaptive particle swarm algorithm, five candidate implant site images (i.e., ROI region of interest) in the quintuple region of the target CBCT image are extracted; then, the peak shape index of the target implant site is used to quantify whether the compatibility between the single alveolar bone and the implant corresponding to the current candidate implant site meets or exceeds the standard compatibility threshold; if so, the current candidate implant site is determined as the target implant site; then, three to five target implant sites are selected from the quintuple region (i.e., target implant sites are selected by analysis and calculation from the quintuple region; if the current target implant sites are three, the subsequent overall evaluation steps (quintuple overall implant bearing evaluation index) are directly executed); see Figure 4-Figure 6 However, the current number of target implant sites is greater than three, for example, 4-5, which may involve many possible combinations of target implant sites (for example, if all five target implant sites meet the requirements, that is, sites 1-5 or sockets 1-5 meet the requirements, it is possible to generate multiple site combinations such as "1-2-3", "1-2-4", "1-2-5", "1-3-4", "1-3-5", "2-3-4", "2-3-5", and "3-4-5"). This is followed by calculating the overall evaluation of multiple target implant combinations (the pentad overall implant load evaluation index) and a boundary area image of the pentad area where the current target implant site is located.

[0072] S30: Pixel analysis is performed based on the target implant site image and the boundary region image of the pentagonal region, and the grayscale texture of the adjacent tooth medullary cavity of the target implant site and the grayscale jump characteristics of the pentagonal region are combined to calculate the overall implant load evaluation index of the pentagonal region (i.e., the overall implant stability is determined by factors such as the grayscale texture of the adjacent tooth medullary cavity and the grayscale jump characteristics of the pentagonal region. However, the stability of the overall implant site is different from the stability of a single implant, so studying the overall implant load evaluation index of the pentagonal region is an important technical approach);

[0073] Then determine whether the implant combination is one type or multiple types according to the current target;

[0074] If the current target implant combination is one, determine whether the overall implant load evaluation index of the quintuple is greater than the minimum overall load threshold. If so, determine that the current target implant combination is the final target implant combination;

[0075] If there are multiple target combination implants, determine whether the corresponding five-body overall implant load evaluation index of each target combination implant is greater than the minimum overall load threshold. If all are greater than the minimum load threshold, establish a numerical list of load evaluation indexes and select the first screening combination processing condition to determine the final target combination implant or determine the final target combination implant according to the second screening combination processing condition. Among them, determining the final target combination implant according to the second screening combination processing condition is one of the important technical means.

[0076] That is, in step S20 of the embodiment of the present application, the suitability of a single alveolar bone is first considered, then the target implant sites that can be selected are screened, and then the possibility of target implant site combinations is analyzed and calculated;

[0077] Then, when step S30 is executed, the target implant site combination possibilities and the pentad implant load evaluation index are comprehensively considered, and the specific three points are selected from the current ones. Finally, the optimal target combined implant combination is output as the final target combined implant combination based on the pentad implant load evaluation index.

[0078] S40: combining the final target implant combination to form a pentad region, calculating the implant length of the target implant site; then determining the initial abutment type based on the overall implant load evaluation index of the pentad region;

[0079] S50: Based on the selected abutment type and implant length, the material usage of the implant length and the abutment type are determined, and a three-dimensional visual implant plan is generated.

[0080] Preferably, as an implementable embodiment, the step S20 of "extracting five candidate implant site images (i.e., ROI image regions or ROI regions of interest) in the quintuple region of the target CBCT image based on an adaptive particle swarm algorithm, and then quantifying whether the compatibility of the single alveolar bone and the implant corresponding to the current candidate implant site meets or exceeds a standard compatibility threshold value by using a peak shape index of the target implant site, and if so, determining the current candidate implant site as the target implant site" includes:

[0081] S21: First, perform particle swarm initialization and adaptive parameter setting: define an initialization particle swarm on the preprocessed quintuplet region image; wherein, define each particle to represent the position of a candidate implantation site and its corresponding adaptive parameters; the adaptive parameters include the position, velocity, and state parameters of the initialization particle;

[0082] It should be noted that a particle swarm is initialized on a preprocessed (denoising, edge enhancement) quintuplet region image to provide a starting point for the search for candidate ROIs. From this starting point, the search begins to locate the initial region of interest and location. In practice, regional scanning and preliminary positioning are performed first, and edge detection is used to locate possible high-contrast areas within the quintuplet region. Specifically, image detection is used to analyze the grayscale uniformity of alveolar bone density and adjacent tooth medullary cavity texture to determine the initial region of interest. Particle initialization is also performed simultaneously. This initialization defines the location of each particle representing a candidate implant site and its corresponding adaptive parameters. The particle's position, velocity, and state parameters are initialized. This velocity is related to the particle position adjustment step size, which is used to dynamically adjust the search range and step size.

[0083] The position of the particle represents the specific position of the candidate implant site within the entire pentagonal region. In this embodiment, the position parameter determines which region in the image the current particle focuses on, which is the specific candidate point in the search space.

[0084] The particle's velocity represents the amplitude and direction of the particle's movement in the search space, i.e., the amount of adjustment corresponding to the ROI position or parameter change. In this embodiment, on the quintuplet region image, the velocity can be understood as the next update step size for the candidate position (e.g., the ROI center position). Generally speaking, there are five dental fossa positions on the quintuplet region image that are the primary candidate regions, but which specific fine-tuning position point on the five dental fossa positions is the most effective ROI image region? If the step size is magnified, the particle's velocity can also be understood as moving from the current dental fossa position to the next dental fossa position. Normally, the particle's velocity is determined by the particle's movement from the current particle position on the quintuplet region image to many other fine-tuning positions. For example, the particle's velocity can be set to move from the current position to a fine-tuning position of 0.5-1.5 mm to the right, or to a fine-tuning position of one-eighth of the current dental fossa width to the right.

[0085] The state parameters of the particle include the particle's current fitness value and the maximum number of iterations.

[0086] S22: performing iterative search and update processing operations based on the adaptive particle swarm; first, defining a local fitness function for each particle, wherein the local fitness function is calculated based on local grayscale value, gradient, texture feature and edge feature.

[0087] The local fitness function is defined for each particle to comprehensively consider the characteristics within the ROI area: ;

[0088] in: : Normalized value of the average gray value in the ROI area; : Normalized value of gradient amplitude in ROI area; : Texture feature normalization value; : normalized value of edge features; α, β, ε and is the corresponding weight coefficient;

[0089] in: : Normalized value of the average gray value in the ROI area. : Normalized value of the gradient amplitude in the ROI area. : Normalized value of texture feature. : Normalized value of edge features (such as edge sharpness). The above coefficients are determined by fitting empirical data and reflect the importance of each feature.

[0090] The average gray value in the ROI region in the above local fitness function (reflects bone tissue density) reflects the bone tissue density;

[0091] The gradient amplitude within the above ROI area (indicating local edge clarity);

[0092] above Texture information (describing the fine structure of tissue within an area);

[0093] above Edge features (reflecting the layered jump in the image).

[0094] For example, α, β, ε, and When the corresponding values ​​are "0.3, 0.4, 0.2, 0.1", and α, β, ε and The sum of the values ​​of is 1, and the corresponding local fitness function is specifically: ;

[0095] Initialize candidate particles; several particles are randomly initialized in the preprocessed quintuplet area.

[0096] For example, the initial position of particle 1 is (80, 120) (image pixel coordinates), the initial position of particle 2 is (90, 130), etc. Each particle is also assigned an initial velocity, for example, the initial velocity of particle 1 is (2, -1).

[0097] Assume particle 1 is at position (80, 120). Within the ROI centered at (80, 120), the extracted normalized feature values ​​are: Inorm=0.7; Gnorm=0.6; Tnorm=0.5; Enorm=0.8. Then calculate the local fitness of the current particle 1.

[0098] in: : Normalized value of the average gray value in the ROI area. : Normalized value of the gradient amplitude in the ROI area. : Normalized value of texture feature. : Normalized value of edge features (such as edge sharpness). The above coefficients are determined by fitting empirical data. If the weights of the coefficients in this embodiment correspond to The corresponding weight is 0.3; The corresponding weight is 0.4; The corresponding weight is 0.2 and The corresponding weight is 0.1.

[0099] The local fitness result then output is F1=0.3×0.7+0.4×0.6+0.2×0.5+0.1×0.8=0.63.

[0100] Assume particle 2 is at position (90, 130); after extracting features, assume the calculated fitness F2 = 0.68, which is the global optimal solution (gbest) in the swarm (assuming the current swarm has only 2 particles). Figure 2 The original CBCT image is acquired, where the implant area to be processed is the maxillary half-mouth area; then Figure 3 This is the effect diagram after particle analysis, in which the two horizontal straight lines at the upper teeth are the directions of the reference step search. Figure 3 Each equilateral triangle in can be schematically represented as a particle; the particle can move and search within each corresponding tooth socket position and in the transverse straight line direction.

[0101] The technical effect achieved in step S22 is based on local optimization, utilizing an adaptive strategy to update the position and velocity of each particle, ensuring that the search process quickly converges to the target region. When the fitness change falls below a preset threshold or reaches the maximum number of iterations, the iteration ends, and the ROI where the particles are currently concentrated, with high local fitness, is saved as the candidate implantation site image. This is the execution goal of step S22: through iterative optimization, the particle swarm is gradually focused on the possible target implantation site region; finally, the peak shape index and fitness formula are calculated for the determined ROI region, followed by quantitative evaluation.

[0102] S23: Perform peak shape index and suitability quantitative evaluation on the determined ROI area to obtain the suitability of the current candidate site:

[0103] The execution goal of the above step S23 is to calculate the compatibility between each candidate ROI region extracted by the adaptive particle swarm algorithm and a single alveolar bone and an implant using a peak shape index.

[0104] Specific operations:

[0105] First, multi-layer slice image analysis: For each candidate implant site, obtain the grayscale value of each pixel (kth pixel) in the jth slice of the ROI area ; Get the lowest gray value in the candidate ROI area ;

[0106] Second, peak index extraction: determine the peak value of the candidate ROI area (such as the maximum grayscale value) as a reference indicator.

[0107] Third, perform the suitability calculation and processing operation:

[0108] ;in: : No. The fitness of each candidate site; : No. The planting site is in Layer slice The gray value of pixels; Layer slices refer to vertical image images of different heights that can be seen and detected through the maxillary alveolar bone as the vertical height changes;

[0109] : The lowest grayscale value in the candidate ROI area (used for normalization or removing background noise);

[0110] : The average gray value of all pixels along the alveolar ridge top in the candidate ROI area;

[0111] : weight coefficient;

[0112] : A very small positive value to prevent division by zero.

[0113] Fourth, fitness evaluation: Calculate the fitness of each candidate site using the above formula to subsequently determine whether it meets the standard fitness threshold.

[0114] S24: Comparison and screening of candidate site fitness thresholds;

[0115] The execution goal of the above step S24 is to compare the fitness of the candidate sites with the preset standard fitness threshold and screen qualified ROI areas.

[0116] Specific operation: For each candidate site, compare its fitness Ai to see if it satisfies Ai ≥ Tadapt (the standard fitness threshold). Candidate sites that meet the criteria are retained, while those that do not are eliminated. Record the number of qualified candidate sites: If the number of candidates is exactly 3, proceed directly to the subsequent overall evaluation step; if the number of candidates is greater than 3 (e.g., 4–5), proceed to the multi-scheme calculation process for candidate combinations.

[0117] S25: Determine the target candidate ROI area and boundary image:

[0118] The execution goal of the above step S25 is to finally determine the target implant sites (3-5) extracted from the quintuplet region and their corresponding boundary region images.

[0119] Specific operation: Combined with the suitability evaluation and the spatial distribution of candidate sites, further screen out 3-5 target implant sites that are balanced and most representative. For each determined target site, extract the boundary area image within a certain range around it at the same time to facilitate a more refined regional evaluation later. Output the target ROI image and its boundary image to provide input for the calculation of the overall implant load evaluation index (i.e., step S30). In the above steps, only if the implant site is greater than the standard suitability threshold can the current candidate implant site be determined as the target implant site and the subsequent step S30 processing operation be performed. Of course, if the total number of the above target implant sites is less than three, the subsequent fuzzy calculation operation of the pentad overall implant load evaluation will be abandoned.

[0120] During the execution of S30, pixel analysis is performed based on the target implant site image and the boundary area image of the quintuplet region, and the grayscale texture of the adjacent tooth medullary cavity of the target implant site and the grayscale jump characteristics of the quintuplet region are combined to calculate the overall implant load evaluation index of the quintuplet, specifically including:

[0121] The detailed steps of the above S30 are expanded from steps S31 to S33, and specifically include:

[0122] S31: Pixel feature extraction is performed on the ROI image of the target implantation site and the corresponding boundary area image.

[0123] When extracting features from the ROI image of the target implant site, key pixel features within the ROI (such as average grayscale, grayscale variance, gradient, and texture direction) are extracted using methods such as local binarization, texture analysis, and edge detection. When extracting boundary features from the ROI image of the target implant site, edge detection is performed on the boundary region image to capture grayscale transition information. Edge sharpness and the grayscale uniformity of the adjacent tooth medullary cavity texture are calculated, with edge sharpness reflecting regional stratification characteristics. During data fusion, the extracted key pixel features within the ROI (edge ​​sharpness) are combined with the grayscale uniformity of the adjacent tooth medullary cavity texture to provide a quantitative basis for subsequent comprehensive evaluation indicators.

[0124] S32: After data fusion, the pentad implant load evaluation index is calculated based on the image data;

[0125] The pixel features obtained in S31 are used to calculate the load evaluation index of the pentagonal implant to reflect the stability of the implant.

[0126] Establish the calculation formula for the evaluation index of the pentad implant load: ;in: It is an evaluation index for the overall implant load;

[0127] The average grayscale value of the target ROI area corresponding to any three target implant sites (reflecting the bone density quality of the overall structure);

[0128] is the average value of edge sharpness in the boundary area corresponding to each of the three target planting sites;

[0129] Grayscale uniformity value of the adjacent tooth medullary cavity texture corresponding to each of the three target implant sites; : Weight coefficient determined by fitting a large amount of experimental data.

[0130] For each candidate target combination (if there are more than three candidate planting sites, various possible three-site planting combinations need to be considered), the above formula is used to calculate the corresponding bearing evaluation index.

[0131] Is the average value of edge sharpness in the boundary area; edge sharpness reflects the intensity and clarity of the grayscale value jump in the boundary area of ​​the image, usually through the gradient change rate. In dental imaging, edge sharpness can reflect the layered clarity of the bone tissue around the teeth. For example: Higher values ​​represent clearer margins: this indicates better structural integrity of the alveolar bone and less bone resorption, which may suggest less severe periodontitis-level resorption. The lower the value, the more blurred the edge: it indicates bone loss and severe alveolar bone absorption (especially severe horizontal absorption of alveolar bone will reduce the sharpness of the edges of most implant sites), suggesting the progression of periodontitis or other inflammatory lesions.

[0132] Teeth are composed of enamel, dentin, and pulp. CBCT images essentially generate grayscale images based on the tissue's absorption of X-rays (density). Denser tissue absorbs more X-rays and appears whiter (higher grayscale values) on CBCT images. Dental tissue with lower density absorbs less X-rays and appears darker (lower grayscale values) on CBCT images. Enamel is the densest and hardest tissue in the body. It appears as the brightest white (highest grayscale value) on CBCT, with sharp, defined boundaries. Dentin is less dense than enamel but significantly denser than pulp. It appears as a uniform light to medium gray (medium-high grayscale value) on CBCT, lying beneath the enamel and forming the main body of the tooth. Dental pulp, primarily composed of loose connective tissue, blood vessels, and nerves, has a very low density (closer to soft tissue). It appears as a clearly visible black or dark gray (lower grayscale value) within the pulp chamber and root canals on CBCT. However, the present embodiment analyzes the characteristics of the alveolar bone at the implant site after the tooth has been extracted. Since the dental pulp is connected to the medullary cavity of the dental socket, when pulpitis spreads to the infection, it will affect the inflammation inside the dental socket. That is, the dental pulp cavity is directly connected to the (medullary cavity) in the alveolar bone through the apical foramen. In pulpitis or apical periodontitis, bacterial toxins can spread through the apical foramen, the periapical tissue, and the medullary cavity, ultimately leading to osteomyelitis. This osteomyelitis can be seen on CBCT.

[0133] The grayscale uniformity value of the adjacent tooth medullary cavity texture corresponding to each of the three target implant sites is calculated. The adjacent tooth medullary cavity texture refers to the location of a portion of the center of the entire implant site. In a normal medullary cavity at this location, the cancellous bone presents a uniform honeycomb-like medium grayscale (grayscale value range: -200~300 HU). However, if this location is infected with osteomyelitis, the uniform medium grayscale area at this location will diminish and become incomplete. Instead, the overall medullary cavity texture grayscale uniformity value will decrease, even to a low level. In addition, the average grayscale value of the adjacent tooth medullary cavity texture will also decrease. In other words, if the grayscale uniformity of the auxiliary adjacent tooth medullary cavity texture (i.e., the low-density abscess cavity further leads to extremely uneven grayscale distribution) is lower than the standard grayscale uniformity threshold, the texture of the corresponding adjacent tooth medullary cavity can also be further judged.

[0134] In the embodiment of the present invention It is the evaluation index of the overall implant load, however, it determines the evaluation index of the overall implant load The average gray value of the ROI area, the average value of the edge sharpness in the boundary area, and the gray uniformity value of the adjacent tooth medullary cavity texture; the overall implant load evaluation index in The reason why we only selected the grayscale uniformity values ​​of the bone marrow cavity texture of the adjacent teeth corresponding to the three target implant sites, instead of selecting the grayscale uniformity values ​​of the bone marrow cavity texture of each implant site, is because such calculation is meaningless, because the three implant sites will eventually undergo dental socket implant surgery, so the height and width of their alveolar bones will be reduced, and the inflammation of the bone marrow cavity will be eliminated and the alveolar bone wear will be treated before the surgery; therefore, it is more meaningful to study the grayscale uniformity values ​​of the bone marrow cavity texture of the adjacent sites, although the inflammation of the bone marrow cavity will also be eliminated, but the alveolar bone wear operation is less; The higher the value, the healthier the adjacent teeth.

[0135] S33: Verification of evaluation index thresholds and determination of target combinations; that is, based on the pentad implant load evaluation index calculated in S32, candidate target combinations are screened and the final implant combination plan is determined. Specific operations:

[0136] First, perform threshold comparison, that is, compare the evaluation index C corresponding to each candidate target combination with the minimum overall carrying threshold Tcarry:

[0137] If the evaluation index C ≥ Tcarry corresponding to the candidate target combination, it is considered that the target combination meets the implant bearing requirements;

[0138] Then determine whether the current target combination implant conjoint is one or more; determine the final target combination implant conjoint according to the five-in-one overall implant load evaluation index; that is, any three target implant sites mentioned above refer to at least three target implant sites that meet the suitability conditions, and the combination method cannot be changed during the same calculation. For example, if there are five target implant sites that meet the conditions, then each time the five-in-one overall implant load evaluation index of the combination is calculated, the corresponding overall implant load evaluation index C is calculated in a fixed combination method; in this way, the corresponding "1-2-3", "1 -2-4", "1-2-5", "1-3-4", "1-3-5", "1-4-5", "2-3-4", "2-3-5", and "3-4-5" to form the values ​​of 9 overall implant load evaluation indicators C; then determine whether these 9 are all greater than Tcarry; for example, assuming that there are only three target implant sites that meet the conditions, such as the site combination of the fine-tuning position of the "1-3-5" tooth socket, then if the calculated overall implant load evaluation indicator C also meets the adjustment greater than Tcarry, then the target combination implant conjoined body will be directly used as the final target combination implant conjoined body;

[0139] Specifically, if there is only one target combination that meets the conditions (i.e., a single target combination implant), processing is performed directly according to the final target combination. However, if there are multiple target combinations that meet the conditions (i.e., multiple plans for multiple target combination implants), a candidate list is created and sorted from high to low according to the evaluation index value.

[0140] Among them, if the current target combination implant conjoint is one, it is judged whether the overall implant load evaluation index of the five-joint is greater than the minimum overall load threshold. If so, the current target combination implant conjoint is determined to be the final target combination implant conjoint;

[0141] Among them, if there are multiple target combination implants at present, it is determined whether the five-body overall implant bearing evaluation index corresponding to each target combination implant is greater than the minimum overall bearing threshold. If all are greater, a numerical arrangement list of bearing evaluation indexes is established and the first screening combination processing condition is selected to determine the final target combination implant (that is, the target combination implant with the largest value in the current numerical arrangement list of bearing evaluation indexes is selected as the final target combination implant) or the final target combination implant is determined according to the second screening combination processing condition.

[0142] Wherein, the second screening combination processing condition is to set three priority screening levels, including a first priority screening level, a second priority screening level and a third priority screening level;

[0143] See also Figure 5 , wherein the first priority screening level is performed according to the planting site "1-3-5" as the first preferred combination;

[0144] See also Figure 6 , the second priority screening level is performed according to the planting site "1-X-5" as the second preferred combination;

[0145] See also Figure 7 , the third priority screening level is executed according to the planting site "2-3-4" as the third preferred combination.

[0146] See also Figure 5-7 , the above Figure 5 - and 7 are schematic diagrams of three combinations of implant positions selected from the horizontal image of the middle layer of the maxillary alveolar bone; wherein Figure 5 The corresponding first priority screening level is the "1-3-5" implant scheme according to the implant site. The researchers analyzed that its uniform distribution of implants is a better five-in-one implant combination scheme; Figure 6 The corresponding second priority screening level is the planting plan according to the planting site "1-X-5"; Figure 7The third priority screening level is the five-unit implant combination plan according to the implant site "2-3-4".

[0147] In general, the above execution steps ensure that the target ROI area is accurately located in the preprocessed image through the adaptive particle swarm algorithm, and at the same time, a detailed peak index quantification formula is used to provide a quantifiable suitability evaluation for each implant site (i.e., single implant condition evaluation) (S21–S25).

[0148] Utilizing the pixel features and data of the image (bone density, stress distribution, grayscale jumps), a comprehensive index is constructed to achieve a quantitative evaluation of the overall implant bearing capacity (i.e., overall implant condition evaluation), providing data support for the selection of multi-target combinations (S31–S33). The final solution is determined, that is, if only one target combination meets the requirements, that combination is directly selected; if there are multiple combinations that meet the conditions, it is necessary to call the screening combination processing conditions in this embodiment and select the one with the largest evaluation index value as the final target implant combination. The final target combination is used as a reference for subsequent further processing operations; therefore, this embodiment not only considers the evaluation of single implant conditions, but also refers to the quantitative evaluation of the overall implant bearing capacity and the determination of the final target combination implant solution when the overall implant bearing capacity is met.

[0149] The technical solution employed in this embodiment of the present invention achieves precise quantitative evaluation of the load-bearing capacity of dental implants through multidimensional data fusion and intelligent evaluation algorithms. This approach offers the following significant benefits: First, it integrates imaging features (grayscale, gradient, texture), biomechanical parameters (bone density, adjacent tooth stress), and boundary characteristics (edge ​​sharpness, grayscale transitions) to construct an evaluation index model encompassing multiple core parameters. By optimizing weight coefficients through an adaptive particle swarm algorithm, the limitations of traditional single-point bone density assessment are overcome, significantly improving the accuracy of overall load-bearing capacity assessment.

[0150] A two-tiered screening approach is also employed. In the first round, unqualified solutions are filtered using the Tcarry threshold. In the second round, a three-level priority strategy ("1-3-5" → "1-X-5" → "2-3-4") is used to optimize the selection of multiple solutions. This technical approach significantly improves solution screening efficiency while ensuring the safety of the selected solutions.

[0151] This technical solution achieves a technological leap from single-point assessment to overall load-bearing analysis through the organic combination of quantitative evaluation and selection combination decision-making, providing reliable decision-making support for complex dental implant surgery and significantly improving the long-term stability of implants.

[0152] In a specific embodiment, during the execution of S40, it is necessary to combine the final target implant combination to form a quintuple region and calculate the implant length of the target implant site;

[0153] The above steps are specifically based on the final particle position and the selected combination formation plan (i.e., implant arrangement and spatial distribution plan selection), and then comprehensively determine the final fine-tuning position of the three target implant sites (i.e., the implant center axis positioning point);

[0154] Then, according to the fine-tuning position of the target implant site (i.e., the center point position), the implant radius is moved to the left according to the center point as the left boundary, and the implant radius is moved to the right according to the center point as the right boundary; the area between the two boundaries is intercepted (as the peri-implant bone volume analysis area) to calculate the thickness of the area (i.e., the vertical bone volume of the alveolar bone), and the thickness is obtained by subtracting the baseline height (i.e., the gingival margin reference plane of the implant restoration) from the lowest point height of the alveolar ridge in the area between the two boundaries to obtain the available bone height; the implant thickness of the other two target implant sites is also obtained by the above method to obtain the available bone height, and then the three target implant sites are jointly calculated to determine their respective lengths to ensure the final height consistency for easy installation of the connecting crown.

[0155] In one embodiment, during the execution of S40, the initial abutment type is determined based on the overall implant load evaluation index of the pentagonal region. The evaluation index C corresponding to each candidate target combination is compared with the minimum overall load threshold Tcarry to evaluate the overall reliability of the candidate target combination. The evaluation index C can also be further used to analyze the selection of abutment materials.

[0156] The specific steps are as follows:

[0157] The abutment suitability evaluation is Class I: the overall implant load evaluation index C is greater than 2.5 times Tcarry;

[0158] The abutment suitability evaluation is level II: the overall implant load evaluation index C is greater than 1.5 times Tcarry and less than 2.5 times Tcarry;

[0159] Abutment suitability is assessed as Class III: The overall implant load assessment index, C, is greater than Tcarry and less than 1.5 times Tcarry. This index, C, is assessed using a graded quantitative approach; a higher C value indicates a higher overall reliability of the candidate combination. In this case, high-performance or lightweight abutment materials can be selected, allowing for a wider range of compatible abutment types. Conversely, if the C value is too low, a composite abutment is preferred.

[0160] For example, when the C value is high, material selection prioritizes high-strength and high-reliability materials, such as pure titanium or titanium alloys. Titanium and its alloys possess high strength and excellent mechanical properties, allowing them to withstand significant masticatory pressures. They also possess excellent biocompatibility, integrating tightly with bone tissue and minimizing rejection. Titanium and titanium alloy abutments are suitable for implant restorations requiring high reliability, particularly in the posterior region, which experiences greater masticatory forces.

[0161] When the C value is moderate, material selection: Zirconia ceramics can be selected. Zirconia ceramics have high strength and good wear resistance, which can meet daily chewing needs.

[0162] When the C value is low, composite materials can be considered, such as a composite abutment with a titanium base and a zirconia overlay. This composite abutment combines the mechanical stability of titanium with the aesthetics of zirconia. In situations where a single material may not meet reliability requirements, composite materials can improve overall reliability by distributing risk.

[0163] Example 2 Further research found that the position selection and update iteration of the particles determine the specific fine-tuning position of the particles within each single tooth socket. Generally speaking, the position between two tooth sockets is not considered for implantation. In some cases (such as when the alveolar bone is severely absorbed), implantation between the tooth sockets can also be considered. Generally speaking, the position selection and update iteration of the particles determine the specific fine-tuning position of the particles within each single tooth socket. However, Example 1 only performs peak shape index and suitability quantitative evaluation based on the determined ROI area to calculate the suitability of the current candidate site, and does not consider other inflammatory absorption conditions.

[0164] Researchers have found that inflammation is one of the important hidden factors affecting the selection of single implant positions. Bone density will decrease due to periodontal inflammation absorption, and both bone height and bone width will be affected. This is also a variable factor that causes bone density to decline.

[0165] Therefore, the technical solution of Example 2 deeply integrates periodontitis parameters into various aspects of particle swarm optimization, thereby realizing the comprehensive prediction and evaluation of the long-term stability of a single implant by combining anatomical characteristics and pathological factors during the implant site selection process.

[0166] During the particle initialization phase, an inflammation signature detection module, or detection method, was added to extract the periodontal inflammation absorption coefficient (p), which is used to measure the attenuation of alveolar bone density. Research has also found that the pentad area typically requires at least three implants to distribute occlusal forces (some plans use four to support the entire mouth) to avoid mechanical complications caused by excessive cantilever length. Implant spacing requirements are also imposed: for example, the distance between adjacent implants must be ≥3mm to prevent overlapping bone resorption.

[0167] Thus, when performing particle position update iteration in the second embodiment of the present invention, not only the selection of a single implant position needs to be considered, but also the distance between adjacent implants needs to be further analyzed and processed. This allows the particle group position selection optimization process to be closer to the actual situation.

[0168] Specifically, the step S20 further includes performing an optimization process after "extracting five images of the candidate implant sites in the quintuple region of the target CBCT image based on the adaptive particle swarm algorithm, and then quantifying whether the compatibility of the single alveolar bone and the implant corresponding to the current candidate implant site meets or exceeds the standard compatibility threshold value by using the peak shape index of the target implant site, and if so, determining the current candidate implant site as the target implant site":

[0169] The newly added first parameter includes the periodontitis inflammatory absorption coefficient P; the newly added second parameter includes the minimum implant spacing limit value D1;

[0170] The first parameter and the second parameter are introduced based on the iterative search and update operation of the adaptive particle swarm; an optimized local fitness function is defined for each particle, taking into account the local gray value, gradient, texture information and edge features; the optimized local fitness function is defined as: Fb=[F×(1− '×P)]× ; ' is the predetermined adjustment coefficient; P is the periodontitis inflammation absorption coefficient;

[0171] is the spacing limit penalty factor, defined as follows:

[0172] If the distances between the current candidate site and its adjacent candidate sites are The minimum distance between implants is limited to D1. ;

[0173] If the distance between any adjacent candidate sites is The minimum distance between implants is limited to D1. .

[0174] Then other execution operations (such as the peak shape index and the quantitative evaluation step of the ROI area and the final determination of the target implantation site and the corresponding ROI boundary area) are the same as those in the first embodiment;

[0175] The following is an improved technical solution that takes into account both the periodontal inflammatory absorption coefficient and the minimum implant spacing limit (i.e., the distance between adjacent implants must be ≥3mm to prevent overlapping bone resorption):

[0176] S211': Particle swarm initialization and adaptive parameter setting: Initialize the particle swarm on the preprocessed quintuple region image, where each particle represents the position of a candidate implantation site and the corresponding adaptive parameters (including initial position, velocity, and state parameters).

[0177] New parameter 1 - periodontitis inflammation absorption coefficient, denoted as P, reflects the absorption of single alveolar bone under inflammatory conditions. P will subsequently play a regulatory role in the calculation of plant suitability. The periodontitis inflammation absorption coefficient cannot be seen through digital images. It can be detected by chemical means to detect the dental plaque of a single tooth. Finally, by calling the preset periodontitis inflammation absorption coefficient and dental plaque comparison table, the periodontitis inflammation absorption coefficient of a single tooth at the tooth socket position corresponding to the current particle's fine-tuning position can be obtained;

[0178] New parameter 2 - minimum implant spacing limit value, denoted as D1: defines a minimum implant spacing value (for example, minimum implant spacing D1 When initializing the candidate site, the spatial position of each particle is recorded. This restriction is then taken into account when determining the positions of adjacent particles to avoid overlapping bone resorption effects caused by placing two adjacent implants too close together.

[0179] S221': Iterative search and update operation based on adaptive particle swarm; defining a local fitness function for each particle, comprehensively considering local grayscale value, gradient, texture information and edge features; the original local fitness function is defined as: ;

[0180] in, : Normalized value of the average gray value in the ROI area; : Normalized value of gradient amplitude in ROI area; : Texture feature normalization value; : Normalized value of edge features (such as edge sharpness);

[0181] After introducing the constraints of the first and second parameters, the local fitness function after fitness optimization is adjusted to:

[0182] Fb=[F×(1− '×P)]× ;

[0183] ' is a predetermined adjustment coefficient used to balance the influence between image features and inflammation absorption factors; F is the value calculated by the original local fitness function; P is the periodontitis inflammation absorption coefficient; is the spacing limit penalty factor;

[0184] If the distances between the current candidate site and its adjacent candidate sites are (i.e. the minimum implant spacing limit value), then the spacing limit penalty factor ; If the distance between any adjacent candidate sites is (i.e. the minimum implant spacing limit value), then the spacing limit penalty factor ( You can set a descending function according to the distance or directly give a larger penalty, and Call to get the preset when a penalty is required value).

[0185] During the above iterative process, each particle adjusts its position and adaptive parameters according to the update formula, and also detects whether the updated positions of adjacent particles meet the minimum implant spacing requirement.

[0186] S231': Quantitative evaluation of peak shape index and fitness of ROI area: For each candidate site in the ROI area, a secondary quantitative evaluation is performed using the local peak shape index to calculate the fitness of the candidate site. .

[0187] Adaptability In the calculation of , in addition to the grayscale, gradient, texture, and edge information that have been considered, the following are also taken into account: the periodontal inflammation absorption coefficient (which reduces the suitability of areas with higher inflammation absorption) and the limiting factor of the distance between adjacent implants (if the site spacing is insufficient, the suitability is additionally penalized).

[0188] Summarize the fitness of all candidate sites Values ​​are used for subsequent screening and comparison.

[0189] S241': Screening and quantity control of candidate sites; for each candidate site, determine its suitability Whether it meets:

[0190] Adaptability ;in, is the standard suitability threshold after adjustment for inflammatory factors and spacing constraints;

[0191] If the candidate site's fitness satisfies At the same time, the distance between the candidate site and the adjacent site is , then the candidate site is retained; otherwise, the candidate sites that do not meet the conditions are eliminated.

[0192] If the number of candidate sites is exactly 3, the system will directly proceed to the subsequent pentad implant load evaluation step; if the number of candidate sites is greater than 3, the system will need to proceed to the multi-scheme overall implant load evaluation step of the candidate combination, and conduct a comprehensive evaluation of different combinations to ensure the optimal load-bearing capacity while meeting anatomical and pathological conditions.

[0193] S251′: Final determination of the target implantation site and the corresponding ROI boundary region: Based on all iterations, suitability evaluation and candidate site screening results, the candidate ROI region of the target implantation site and its corresponding boundary region image are finally determined from the quintuple region.

[0194] The final implant site must meet the required image features (grayscale, gradient, texture, edges, etc.), while also providing sufficient bone support under the influence of periodontal inflammatory resorption. The spacing between adjacent implants must meet the constraint of no less than 3 mm to ensure that the long-term stability of the implant is not compromised by cumulative bone resorption during actual use. The alveolar bone, also known as the alveolar process, is the portion of the maxillary and mandibular bones that surrounds and supports the tooth roots. Alveolar bone resorption is a key pathological change in periodontitis. This loss of alveolar bone leads to loss of supporting tissue and gradual tooth mobility. Research has found that the alveolar bone is the most metabolically active and remodeling component of the periodontal tissue and the entire skeletal system. Under normal physiological conditions, alveolar bone resorption and regeneration are balanced, and alveolar bone height remains constant. When bone resorption increases, bone regeneration decreases, or both occur, bone loss occurs, reducing alveolar bone height. Therefore, the improved technical solutions studied in the embodiments of this application take into account the periodontal inflammatory resorption coefficient P and the minimum spacing limit between adjacent implants. The final output plan will provide data basis for subsequent clinical implant load evaluation, implant surgery preparation and implant consumables.

[0195] Example 3 In this embodiment, after determining the final target combined implant combination according to the second screening combination processing condition, the method further includes:

[0196] Further testing of the implant sites in the current ultimate target implant combination;

[0197] Determine whether the adjacent teeth at the current implant site are tilted;

[0198] If it is determined that the adjacent teeth at the current implant site are tilted, it is determined that the current final target combined implant joint encroaches on the implant space;

[0199] Filter out the current final target combination implant for further orthodontic or grinding treatment.

[0200] Example 4 like Figure 8 As shown, on the other hand, this third embodiment, based on the artificial intelligence oral surgery consumables usage evaluation and optimization method provided in the first embodiment of the invention, also provides a computer storage medium 1140 (hereinafter referred to as the storage medium). This is a schematic diagram of the computer storage medium structure framework provided in the third embodiment of the invention, which includes:

[0201] Memory 1130, for storing computer programs;

[0202] The communication interface 1120 is used to connect the memory 1130 to the processor 1110;

[0203] Processor 1110 is used to execute a computer program to implement an artificial intelligence oral surgery consumables usage evaluation optimization method involved in Example 1 disclosed as a combination of any of the above embodiments. It will be understood that the embodiments described herein can be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits, digital signal processors, digital signal processing devices, programmable logic devices, field programmable gate arrays, general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in this application, or a combination thereof. For software implementation, the technology described herein can be implemented by a unit that performs the functions described herein. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0204] To sum up, the example of the present invention proposes an artificial intelligence method for evaluating and optimizing the use of oral surgical consumables. This method adopts technical factors such as denoising, edge enhancement, peak index quantification, and pixel analysis in single image recognition; in the overall pentad area evaluation and implant combination selection, it combines image processing optimization, algorithm application, comprehensive evaluation indicators, absorption parameter determination and other technical points to achieve overall stability evaluation, effectively improving the quality of implant plan selection for the pentad implant during surgery, and laying a solid foundation for subsequent implant material selection.

[0205] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. A person skilled in the art may modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for calculating and predicting the amount of implant material used based on a particle swarm image algorithm for oral digital modeling, characterized in that: The following steps are included: Acquire the patient's oral 3D imaging data, including the grayscale texture of the adjacent tooth medullary cavity and the grayscale jump characteristics of the pentad area; Denoising and edge enhancement are performed on the three-dimensional image data to obtain a target CBCT image; based on an adaptive particle swarm algorithm, five images of candidate implant sites in the quintuplicate region of the target CBCT image are extracted; then, using a peak shape index of the target implant site, the compatibility between the single alveolar bone and the implant corresponding to the current candidate implant site is quantified to be greater than a standard compatibility threshold; if so, the current candidate implant site is determined as the target implant site; then, three to five target implant sites and boundary region images of the quintuplicate region where the current target implant site is located are selected from the quintuplicate region; Perform pixel point analysis based on the target implant site image in the quintuplet area and the boundary area image of the quintuplet area, and calculate the quintuplet overall implant load evaluation index based on the grayscale texture of the adjacent tooth medullary cavity of the target implant site and the grayscale jump characteristics of the quintuplet area; then determine whether the current target combination implant is one or more; and determine the final target combination implant based on the quintuplet overall implant load evaluation index; Combine the final target implant combination to form a pentad area, and calculate the implant length of the target implant site; Then, the initial abutment type is determined based on the overall implant load evaluation index of the pentagonal area; then, based on the selected abutment type and implant length, the material usage and abutment type of the implant length are determined, and a three-dimensional visual implant plan is generated.

2. The method according to claim 1, characterized in that The adaptive particle swarm algorithm is used to extract five candidate implant site images in the quintuple region of the target CBCT image, and then quantify whether the compatibility between the single alveolar bone and the implant corresponding to the current candidate implant site meets or exceeds the standard compatibility threshold through the peak shape index of the target implant site. If so, the current candidate implant site is determined as the target implant site, specifically including: First, particle swarm initialization and adaptive parameter setting are performed: an initialization particle swarm is defined on the preprocessed quintuplet region image; each particle is defined to represent the position of a candidate implantation site and its corresponding adaptive parameters; the adaptive parameters include the position, velocity, and state parameters of the initialization particle; During initialization, a local fitness function is defined for each particle, which is calculated based on the normalized value of the average grayscale value in the ROI area, the normalized value of the gradient amplitude in the ROI area, the normalized value of the texture feature, and the normalized value of the edge feature; Perform peak shape index and fitness quantitative evaluation on the determined ROI area to calculate the fitness Ai of the current candidate site; summarize and calculate the fitness Ai of each candidate site, and judge its relationship with the fitness threshold of the Tadapt standard to screen candidate sites; For each candidate site, compare its suitability Ai to see if it meets the standard suitability threshold of Ai ≥ Tadapt; retain the candidate sites that meet the conditions, and eliminate those that do not; record the number of qualified candidate sites: if the number of candidates is exactly 3, directly proceed to the subsequent evaluation step of the pentad integral implant load-bearing evaluation index; if the number of candidates is greater than 3, proceed to the evaluation step of the pentad integral implant load-bearing evaluation index of the multiple schemes of the candidate combination.

3. Finally determine the target candidate ROI area of ​​the target implantation site extracted from the quintuple region and its corresponding boundary area image.

4. The method according to claim 2, characterized in that The step of performing a quantitative evaluation of the peak shape index and the suitability metric on the determined ROI region to obtain the suitability Ai of the current candidate site specifically includes the following steps: For each candidate implant site, obtain the grayscale value of each pixel in the jth slice of the ROI area ; Get the lowest grayscale value in the candidate ROI area; determine the peak value of the candidate ROI area As a reference indicator; Perform fitness calculation processing operations: ； in: : No. The fitness of each candidate site; : No. The planting site is in Layer slice The grayscale value of each pixel; : The lowest gray value in the candidate ROI area; : The average gray value of all pixels along the alveolar ridge top in the candidate ROI area; : is the weight coefficient; : A very small positive value to prevent division by zero.

5. The method according to claim 3, characterized in that The method comprises performing pixel point analysis based on the target implant site image in the quintuplet area and the boundary area image of the quintuplet area, and calculating the quintuplet overall implant load evaluation index in combination with the grayscale texture of the adjacent tooth medullary cavity of the target implant site and the grayscale jump characteristics of the quintuplet area; then judging whether the current target combined implant conjoint is one or more than one; and determining the final target combined implant conjoint based on the quintuplet overall implant load evaluation index, specifically including: Pixel feature extraction is performed on the ROI image of the target implant site and the corresponding boundary area image. When extracting features from the ROI image of the target implant site, the average grayscale value of the pixel features within the ROI is extracted. When extracting boundary features from the ROI image of the target implant site, the edge sharpness and grayscale uniformity of the adjacent tooth medullary cavity texture are extracted based on the boundary grayscale jump information. The evaluation index of the pentad implant load was calculated based on the average gray value of the pixel features within the ROI, the edge sharpness, and the gray uniformity of the adjacent tooth medullary cavity texture. ； in: It is an evaluation index for the overall implant load; is the average gray value of the target ROI area corresponding to any three target implantation sites; is the average value of edge sharpness in the boundary area corresponding to each of the three target planting sites; Grayscale uniformity value of the adjacent tooth medullary cavity texture corresponding to each of the three target implant sites; is the corresponding weight coefficient; Based on the calculated pentad implant load evaluation index, candidate target combinations are screened and the final implant combination plan is determined. The specific operations are as follows: First, a threshold comparison is performed, that is, the evaluation index C corresponding to each candidate target combination is compared with the minimum overall load threshold Tcarry: if the evaluation index C corresponding to the candidate target combination ≥ Tcarry, then the target combination is considered to meet the implant load requirements; Then, it is determined whether the current target combined implant combination is one or more than one; and the final target combined implant combination is determined based on the five-unit overall implant load evaluation index; If the current target implant combination is one, determine whether the overall implant load evaluation index of the quintuple is greater than the minimum overall load threshold. If so, determine that the current target implant combination is the final target implant combination; If there are multiple target combination implants at present, determine whether the five-body overall implant load evaluation index corresponding to each target combination implant is greater than the minimum overall load threshold. If all are greater, establish a numerical arrangement list of load evaluation indexes and select the first screening combination processing condition to determine the final target combination implant or determine the final target combination implant according to the second screening combination processing condition.

6. The method according to claim 4, characterized in that The first screening combination processing condition is to select the target combination implant conjoint with the largest value in the current load evaluation index value arrangement list as the final target combination implant conjoint.

7. The method according to claim 5, characterized in that The second screening combination processing condition is to set three priority screening levels, including a first priority screening level, a second priority screening level, and a third priority screening level; Among them, the first priority screening level is to execute according to the planting site "1-3-5" as the first preferred combination; the second priority screening level is to execute according to the planting site "1-X-5" as the second preferred combination; the third priority screening level is to execute according to the planting site "2-3-4" as the third preferred combination.

8. The method according to claim 3, characterized in that After "extracting five candidate implant site images in the quintuple region of the target CBCT image based on an adaptive particle swarm algorithm, and then quantifying whether the compatibility of the single alveolar bone and the implant corresponding to the current candidate implant site meets or exceeds a standard compatibility threshold using the peak shape index of the target implant site, and if so, determining the current candidate implant site as the target implant site", the following optimization processing operations are also performed: The newly added first parameter includes the periodontitis inflammatory absorption coefficient P; the newly added second parameter includes the minimum implant spacing limit value D1; The first parameter and the second parameter are introduced based on the iterative search and update operation of the adaptive particle swarm; an optimized local fitness function is defined for each particle, taking into account the local gray value, gradient, texture information and edge features; the optimized local fitness function is defined as: Fb=[F×(1− '×P)]× ; ' is the predetermined adjustment coefficient; P is the periodontitis inflammation absorption coefficient; is the spacing limit penalty factor, defined as follows: If the distances between the current candidate site and its adjacent candidate sites are The minimum distance between implants is limited to D1. ; If the distance between any adjacent candidate sites is The minimum distance between implants is limited to D1. .

9. The method according to claim 7, characterized in that After determining the final target combined implant conjoint according to the second screening combined implant processing condition, the method further includes: further detecting the implant sites in the current final target combined implant conjoint; Determine whether the adjacent teeth at the current implant site are tilted; If it is determined that the adjacent teeth at the current implant site are tilted, it is determined that the current final target combined implant joint encroaches on the implant space; Filter out the current final target combination implant for further orthodontic or grinding treatment.

10. A storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the artificial intelligence oral surgery consumables usage evaluation and optimization method according to any one of claims 1 to 8.

Images