An AI generation system for a tooth invisible correction treatment plan

Abstract

The present application relates to the technical field of medical data mining, in particular to a tooth invisible correction treatment scheme AI generation system, first, based on the similarity between historical tooth correction patients in the tooth deformity symptom sequence representing the tooth deformity symptom characteristics and the physiological feature data sequence representing the physiological data characteristics, similar patient clustering clusters are divided; then based on the treatment response characteristics represented by the planning deviation of the actual tooth movement distance of each historical tooth correction patient and the complication frequency distribution, the treatment comprehensive response is determined; finally, taking the sample reference degree represented by the number of samples in the similar patient clustering cluster where the tooth correction patient to be analyzed is located as the weight, combining the feature deviation distance and the treatment comprehensive response, the tooth invisible correction treatment scheme of the tooth correction patient to be analyzed is comprehensively determined, so that the accuracy of the obtained tooth invisible correction treatment scheme is higher.

Description

Technical Field

[0001] This invention relates to the field of medical data mining technology, specifically to an AI-generated system for invisible orthodontic treatment plans. Background Technology

[0002] Existing technologies typically determine the type of malocclusion in a patient and then use the most frequently used orthodontic treatment plan in historical data for that type of malocclusion as the corresponding invisible orthodontic treatment plan. However, this method relies too heavily on a single dimension of morphological classification and does not take into account the differences between individual patients in terms of biophysiology, dental malocclusion symptoms, etc., resulting in low accuracy of the determined invisible orthodontic treatment plan. Summary of the Invention

[0003] To address the low accuracy of existing technologies that determine invisible orthodontic treatment plans for patients based on malocclusion type analysis, this application aims to provide an AI-based invisible orthodontic treatment plan generation system. The specific technical solution adopted is as follows:

[0004] The first aspect of this application provides an AI-generated system for invisible orthodontic treatment plans, including:

[0005] The data acquisition and preprocessing module is used to obtain the physiological characteristic data sequence, dental malocclusion symptom sequence, actual tooth movement distance and all complications of each historical orthodontic patient from the medical database; wherein the historical orthodontic treatment plan of each historical orthodontic patient is the same in all the medical records of all visits.

[0006] The first determination module is used to determine the corresponding feature deviation distance based on the similarity of the dental malocclusion symptom sequence and physiological characteristic data sequence between each historical orthodontic patient and other orthodontic patients; to perform cluster analysis based on the feature deviation distance between historical orthodontic patients to determine all similar patient clusters; and to determine the corresponding sample reference degree based on the number of historical orthodontic patients in each similar patient cluster.

[0007] The second determination module is used to determine the corresponding overall treatment response based on the planned deviation of the actual tooth movement distance and the distribution of complication frequency in the treatment records of each historical orthodontic patient throughout all visits.

[0008] The treatment plan determination module is used to determine the invisible orthodontic treatment plan for the patient to be analyzed within the similar patient clusters in which the patient to be analyzed belongs, based on the referenceability of the corresponding samples and the characteristic deviation distance between each historical orthodontic patient and the patient to be analyzed.

[0009] Furthermore, the process of obtaining the feature deviation distance includes:

[0010] Each historical orthodontic patient was selected as the target orthodontic patient, and other historical orthodontic patients were selected as reference orthodontic patients.

[0011] The similarity of dental malocclusion symptoms is determined based on the Jaccard coefficient between the dental malocclusion symptom sequence of the target orthodontic patient and the dental malocclusion symptom sequence of each reference orthodontic patient.

[0012] The Euclidean distance between the physiological characteristic data sequence of the target orthodontic patient and the physiological characteristic data sequence of each reference orthodontic patient is normalized to obtain the corresponding physiological characteristic deviation.

[0013] The feature deviation distance between the target orthodontic patient and each reference orthodontic patient is determined by multiplying the negative correlation mapping value of the similarity of the dental malocclusion symptoms with the deviation of the physiological characteristics.

[0014] Furthermore, the process of obtaining the clusters of similar patients includes:

[0015] Based on the feature deviation distance, the K-means clustering algorithm was used to perform cluster analysis on all historical orthodontic patients to determine all similar patient clusters.

[0016] Furthermore, the process of obtaining the referenceability of the sample includes:

[0017] The number of patients with historical orthodontic treatment in each similar patient cluster is normalized to determine the reference value of the corresponding sample.

[0018] Furthermore, the process of obtaining the overall treatment response includes:

[0019] The orthodontic achievement rate is determined based on the overall deviation between the actual tooth movement distance and the corresponding planned tooth movement distance in the treatment records of each patient throughout all visits.

[0020] The indirect response performance of orthodontics was determined based on the frequency distribution of complications in the treatment records of each patient throughout all visits.

[0021] The corresponding overall treatment response is determined by multiplying the orthodontic achievement rate with the indirect response performance of the orthodontic treatment.

[0022] Furthermore, the process of obtaining the orthodontic achievement rate includes:

[0023] The difference between the actual tooth movement distance and the corresponding planned tooth movement distance in the treatment records of each historical orthodontic patient at each visit was normalized to determine the corresponding local orthodontic achievement; the corresponding orthodontic achievement was determined based on the mean of the local orthodontic achievement in the treatment records of each historical orthodontic patient across all visits.

[0024] Furthermore, the process of obtaining the corrective indirect response performance includes:

[0025] The frequency of complications for each historical orthodontic patient was arranged chronologically from all treatment records and then curve-fitted to determine the complication frequency curve. The mean of the incisional slope corresponding to all treatments on the complication frequency curve was negatively correlated to determine the trend value of complication changes.

[0026] The overall complication severity was determined based on the mean frequency of complications in all treatment records of each historical orthodontic patient.

[0027] The orthodontic indirect response performance of each patient with a history of orthodontic treatment is determined by multiplying the negative correlation mapping value of the overall performance of the complications with the trend value of the complications.

[0028] Furthermore, the process of obtaining the invisible orthodontic treatment plan includes:

[0029] Based on the referenceability of samples in the similar patient clusters to which the orthodontic patient to be analyzed belongs, a first reference weight is determined; the first reference weight is negatively correlated to determine a second reference weight; the sum of the first reference weight and the second reference weight is 1; historical orthodontic patients in the similar patient clusters to which the orthodontic patient to be analyzed belongs are used as reference patients.

[0030] The weighted response of each reference patient is determined by multiplying the normalized value of the overall treatment response of each reference patient with the first reference weight; the weighted similarity of each reference patient is determined by the second reference weight and the feature deviation distance between each reference patient and the orthodontic patient to be analyzed.

[0031] The fit optimization degree of each reference patient is determined based on the sum of the weighted similarity and the weighted responsiveness.

[0032] Based on the fit optimization and the historical orthodontic treatment plans of each reference patient, invisible orthodontic treatment plans for patients to be analyzed were selected.

[0033] Furthermore, the process of obtaining the weighted similarity includes:

[0034] The feature deviation distance between each reference patient and the orthodontic patient to be analyzed is negatively correlated to determine the corresponding feature similarity; the weighted similarity of each reference patient is determined based on the product of the feature similarity and the second reference weight.

[0035] Furthermore, the process of selecting invisible orthodontic treatment plans for the patients to be analyzed based on the fit optimization degree and the historical orthodontic treatment plans of each reference patient includes:

[0036] The historical orthodontic treatment plan of the reference patient with the highest fit and optimization degree will be used as the invisible orthodontic treatment plan for the patient to be analyzed.

[0037] Secondly, this application provides a computer device including a memory and a processor. The memory is used to store computer program code, and the processor is used to call and run the computer program code from the memory to execute a system as described in the first aspect of this application or any embodiment of the first aspect.

[0038] Thirdly, this application provides a computer program product, which includes computer program code that, when executed, performs a system as described in the first aspect of this application or any embodiment thereof.

[0039] Fourthly, this application provides a computer-readable storage medium that stores computer program code, which, when executed, performs a system as described in the first aspect of this application or any embodiment thereof.

[0040] This application has the following beneficial effects:

[0041] This application first divides patients into similar patient clusters based on the similarity of dental malformation symptom sequences and physiological feature data sequences among patients with a history of orthodontic treatment. Then, it determines the overall treatment response based on the treatment response characteristics represented by the actual tooth movement distance of each patient with a history of orthodontic treatment and the distribution of complication frequency. Finally, it comprehensively determines the invisible orthodontic treatment plan for the patient with a history of orthodontic treatment by using the referenceability of the sample number in the similar patient cluster to which the patient to be analyzed belongs as a weight, combined with the feature deviation distance and the overall treatment response, thus making the obtained invisible orthodontic treatment plan more accurate. Attached Figure Description

[0042] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a structural diagram of an AI-generated invisible orthodontic treatment plan according to an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of a computer device structure provided in one embodiment of the present invention. Detailed Implementation

[0045] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an AI-generated invisible orthodontic treatment plan system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment, and specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature.

[0046] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0047] The following description, in conjunction with the accompanying drawings, details the specific scheme of the AI-generated invisible orthodontic treatment plan system provided by this invention.

[0048] This application provides an AI-generated system for invisible orthodontic treatment plans. Please refer to [link / reference]. Figure 1 The diagram shows a structural diagram of an AI-generated system for invisible orthodontic treatment plans according to an embodiment of the present invention. The system includes: a data acquisition and preprocessing module 101, a first determination module 102, a second determination module 103, and a plan determination module 104.

[0049] The data acquisition and preprocessing module 101 is used to obtain the physiological characteristic data sequence, dental malocclusion symptom sequence, actual tooth movement distance and all complications of each historical orthodontic patient in the medical database; wherein the historical orthodontic treatment plan of each historical orthodontic patient is the same in all the medical records of all visits.

[0050] Quantitative data values ​​for each physiological characteristic dimension are collected for each historical orthodontic patient. The quantitative data values ​​for all physiological characteristic dimensions are arranged sequentially according to the collection order to determine the physiological characteristic data sequence for each historical orthodontic patient. The order of the physiological characteristic dimensions is the same across all historical orthodontic patients' physiological characteristic data sequences. In one specific implementation of this invention, the physiological characteristic dimensions include bone density, age, average alveolar bone height, average alveolar bone velocity, average gingival thickness, and average root length, which can be adjusted according to the specific implementation environment. It should be noted that the quantitative data values ​​for all corresponding physiological characteristic dimensions in this embodiment are normalized values, which reduces the influence of dimensions on the calculation process; this will not be further elaborated upon here. Furthermore, it should be noted that historical orthodontic patients in this embodiment are those who have had more than two treatment visits and whose teeth have been successfully corrected under the corresponding historical orthodontic treatment plan; this will not be further elaborated upon here.

[0051] For each patient's history of orthodontic treatment, the initial consultation record was segmented using the jieba word segmentation method, and then keywords were extracted using the TF-IDF algorithm to identify all corresponding keywords, such as those related to malocclusion types, including crowded teeth, protruding arch, deep overbite, deep overjet, receding chin, underbite, facial asymmetry, and edge-to-edge malocclusion. The sequence of all keywords for each patient's history of orthodontic treatment was then arranged to form the dental malocclusion symptom sequence. It should be noted that both the jieba word segmentation method and the TF-IDF algorithm are well-known techniques in the art, and implementers can adjust the word segmentation method and keyword extraction algorithm according to the specific implementation environment. Furthermore, the consultation record in this embodiment refers to the content recorded by the doctor during the consultation, and no further limitations or elaborations are made here. It should also be noted that the order of the dental malocclusion symptom sequence does not affect subsequent analysis and can be set independently. This embodiment uses the keyword extraction order for arrangement, which will not be further elaborated here.

[0052] In the process of invisible orthodontic treatment, the treatment record serves as a periodic monitoring of dental data. By using the 3D jaw model at each treatment and the 3D jaw model at the previous treatment, the planned and actual tooth movement distances recorded at each treatment can be determined, thereby assessing the orthodontic progress. Furthermore, complications related to orthodontic treatment, such as root resorption, gingival recession, and attachment loss, are extracted from the medical database for each historical orthodontic patient's treatment records at each treatment, and will not be further elaborated here. The 3D jaw model can be determined using a standard intraoral scanner, which will not be further elaborated here. It should be noted that, to ensure the completeness of the embodiment, the planned and actual tooth movement distances at the first treatment in this embodiment are both set to 0, which will not be further elaborated here.

[0053] The first determining module 102 is used to determine the corresponding feature deviation distance based on the similarity of the dental malocclusion symptom sequence and physiological characteristic data sequence between each historical orthodontic patient and other orthodontic patients; to perform cluster analysis based on the feature deviation distance between historical orthodontic patients to determine all similar patient clusters; and to determine the corresponding sample reference degree based on the number of historical orthodontic patients in each similar patient cluster.

[0054] Considering the limitations of single-dimensional similarity analysis, it is necessary to combine multi-dimensional data analysis when performing similarity analysis on patients with a history of orthodontic treatment. Since both physiological feature data sequences and dental malocclusion symptom sequences represent data features under multiple dimensions, this embodiment of the invention further determines the corresponding feature deviation distance based on the similarity of dental malocclusion symptom sequences and physiological feature data sequences between each patient with a history of orthodontic treatment and other patients with orthodontic treatment. The smaller the feature deviation distance, the more similar the two patients with a history of orthodontic treatment are in terms of physiological features and dental malocclusion symptom features.

[0055] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the feature deviation distance includes:

[0056] Each patient with a history of orthodontic treatment was designated as the target patient, and other patients with history of orthodontic treatment were designated as reference patients. The similarity of dental malocclusion symptoms was determined based on the Jaccard coefficient between the sequence of dental malocclusion symptoms of the target patient and each reference patient. According to the definition of the Jaccard coefficient, a higher similarity indicates a greater overlap between the sequence of dental malocclusion symptoms of the target patient and each reference patient, a more similar distribution of keywords representing the dental malocclusion symptoms, and thus a higher similarity in dental malocclusion symptoms between the two patients. It should be noted that the calculation of the Jaccard coefficient is a technique well-known to those skilled in the art and will not be further limited or elaborated upon here.

[0057] The physiological characteristic data sequence of the target orthodontic patient is normalized with the physiological characteristic data sequence of each reference orthodontic patient to determine the corresponding physiological characteristic deviation. Since the physiological characteristic data sequence in this embodiment is a sequence corresponding to quantized data, the corresponding physiological characteristic deviation can be directly determined by calculating the Euclidean distance between the sequences. The smaller the Euclidean distance, the smaller the physiological characteristic deviation, and the higher the similarity in physiological characteristics between the two patients. In another specific implementation of this embodiment, the corresponding physiological characteristic deviation can also be determined by normalizing the cosine similarity between the physiological characteristic data sequence of the target orthodontic patient and the physiological characteristic data sequence of each reference orthodontic patient; this will not be elaborated further here.

[0058] Furthermore, by combining the correlation between similarity of dental malocclusion symptoms and deviation of physiological characteristics with similarity, the feature deviation distance between the target orthodontic patient and each reference orthodontic patient is determined based on the product of the negative correlation mapping value of similarity of dental malocclusion symptoms and deviation of physiological characteristics. The smaller the feature deviation distance, the higher the similarity between the target orthodontic patient and each reference orthodontic patient, and the more likely they are to belong to the same patient cluster.

[0059] In one specific implementation of this invention, the calculation process of the feature deviation distance is expressed by the formula: ;in, For target orthodontic patients With the corresponding first The characteristic deviation distance between reference orthodontic patients; For target orthodontic patients The sequence of dental malocclusion symptoms and the corresponding first The Jaccard coefficient between sequences of dental malocclusion symptoms of reference orthodontic patients, which is the similarity of dental malocclusion symptoms; For target orthodontic patients Physiological characteristic data sequence and corresponding first Euclidean distance between physiological characteristic data sequences of reference orthodontic patients; It is a minimum-maximum normalization function, and the normalization method can be adjusted according to the specific implementation environment; For target orthodontic patients With the corresponding first The degree of deviation in physiological characteristics between reference orthodontic patients.

[0060] Based on the similarity characteristics exhibited by the feature deviation distance, further cluster analysis is performed according to the feature deviation distance between historical orthodontic patients. All historical orthodontic patients are divided into various similar patient clusters, ensuring that the physiological data and dental malocclusion symptom-related data of all patients in each similar patient cluster exhibit highly similar characteristics. This allows for more accurate analysis based on group characteristics when subsequently selecting invisible orthodontic treatment plans for the patients to be analyzed. It should be noted that the physiological characteristic data sequence and dental malocclusion symptom sequence of the patients to be analyzed are pre-acquired data in this embodiment of the invention. The similar patient clusters are determined by calculating the feature deviation distance between the patients to be analyzed and each historical orthodontic patient; further details are omitted here.

[0061] Preferably, in a specific implementation of this invention, the process of obtaining similar patient clusters includes: performing cluster analysis on all historical orthodontic patients based on feature deviation distance using the K-means clustering algorithm to determine all similar patient clusters; in a specific implementation of this invention, the K value of the K-means clustering algorithm is determined using the elbow method. It should be noted that the K-means clustering algorithm and the elbow method are techniques well-known to those skilled in the art, and can be adjusted or other clustering algorithms can be used according to the specific implementation environment; further limitations and descriptions are not provided here.

[0062] For each cluster of similar patients, the larger the number of patients with a history of orthodontic treatment, the more patients within that cluster share similar multi-dimensional characteristics, resulting in a broader range of data samples for reference. This allows for a focus on determining the treatment plan for invisible orthodontics based on the treatment effects of the patients' past orthodontic treatments within the cluster, ensuring better orthodontic performance for the patients. Conversely, fewer data samples with reference value necessitate identifying patients within clusters with similar multi-dimensional characteristics who are closer to the patients to be analyzed in terms of their multi-dimensional features. Therefore, the reference value of the samples is further determined based on the number of patients with a history of orthodontic treatment in each cluster, allowing for subsequent analysis based on the magnitude of the reference value. In one specific implementation of this invention, the process of obtaining the reference value of the samples includes: normalizing the number of patients with a history of orthodontic treatment in each cluster to determine the corresponding reference value of the samples. The normalization method for the number of historical orthodontic patients in each similar patient cluster is min-max normalization. The normalization method can be adjusted according to the specific implementation environment, which will not be elaborated further here.

[0063] The second determining module 103 is used to determine the corresponding overall treatment response based on the planned deviation of the actual tooth movement distance and the distribution of complication frequency in the treatment records of each historical orthodontic patient throughout all visits.

[0064] For each patient with a history of orthodontic treatment, the closer the tooth movement distance during treatment is to the target value and the lower the frequency of complications, the more ideal the treatment effect. Therefore, this embodiment of the invention further determines the corresponding comprehensive treatment response based on the planned deviation of the actual tooth movement distance and the distribution of complication frequency in the treatment records of each patient with a history of orthodontic treatment across all visits. A higher comprehensive treatment response indicates a better orthodontic effect for the corresponding patient with a history of orthodontic treatment. Therefore, in clusters of similar patients with a large number of patients with a history of orthodontic treatment, the determination of the invisible orthodontic treatment plan should focus more on the comprehensive treatment response, which characterizes the treatment effect of each patient's previous orthodontic treatment plan. Thus, this embodiment of the invention determines the corresponding comprehensive treatment response based on the planned deviation of the actual tooth movement distance and the distribution of complication frequency in the treatment records of each patient with a history of orthodontic treatment across all visits.

[0065] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the overall treatment response includes:

[0066] Based on the overall deviation between the actual tooth movement distance and the corresponding planned tooth movement distance in the treatment records of each historical orthodontic patient across all visits, the corresponding orthodontic achievement is determined. Specifically, the difference between the actual tooth movement distance and the corresponding planned tooth movement distance in the treatment records of each historical orthodontic patient at each visit is normalized to determine the corresponding local orthodontic achievement.

[0067] For each patient with a history of orthodontic treatment, the closer the actual tooth movement distance recorded at each visit is to the planned tooth movement distance set in the treatment plan, that is, the greater the local orthodontic achievement, the better the orthodontic treatment is considered. Therefore, the corresponding orthodontic achievement is determined based on the average local orthodontic achievement of each patient across all visits. A higher orthodontic achievement indicates a better treatment outcome for the corresponding orthodontic treatment plan, that is, a higher overall treatment response.

[0068] Based on the frequency distribution of complications in the treatment records of each historical orthodontic patient across all visits, the orthodontic indirect response performance was determined. Specifically, the process of obtaining the orthodontic indirect response performance included:

[0069] The frequency of complications for each historical orthodontic patient was arranged chronologically in all treatment records and then curve-fitted to determine the complication frequency curve. The mean incisional slope corresponding to all treatments on the complication frequency curve was negatively correlated to determine the trend value of complication changes. Based on the mean frequency of complications for each historical orthodontic patient in all treatment records, the overall complication performance was determined.

[0070] For each patient with a history of orthodontic treatment, the lower the frequency of complications as the treatment progresses and the lower the overall frequency of complications, the better the corresponding orthodontic treatment plan is at suppressing complications. Therefore, the larger the complication trend value and the smaller the overall complication performance calculated in this embodiment of the invention, the better the treatment effect of the corresponding orthodontic treatment plan, that is, the greater the overall treatment response. Therefore, by further combining the correlation, the indirect response performance of each patient with a history of orthodontic treatment is determined based on the product between the complication trend value and the overall complication performance, so that the greater the indirect response performance, the greater the overall treatment response.

[0071] In one specific implementation of this invention, the process of obtaining the corrected indirect response performance is expressed by the following formula: ;in, For patients with a history of orthodontic treatment The degree of indirect response performance in correction; For patients with a history of orthodontic treatment The mean of the incision slope corresponding to all visits on the complication frequency curve; For patients with a history of orthodontic treatment The trend value of complication changes; For patients with a history of orthodontic treatment The mean frequency of complications in all medical records of all visits, i.e. the overall complication performance; since the slope of the tangent line under the decay trend of the complication frequency curve is negative, the negative correlation mapping process in this embodiment of the invention is to normalize the negative of the mean of the corresponding tangent slope.

[0072] Finally, based on the correlation, the overall treatment response is determined by multiplying the orthodontic achievement rate with the indirect response performance. A higher overall treatment response indicates a better treatment outcome for the corresponding historical orthodontic treatment plan when dealing with the same patient. In one specific implementation of this invention, the process of obtaining the overall treatment response is expressed by the following formula: ;in, For patients with a history of orthodontic treatment The overall response rate of treatment; For patients with a history of orthodontic treatment Number of medical visits; For patients with a history of orthodontic treatment In the The actual distance the teeth moved as recorded in the medical record of each visit; For patients with a history of orthodontic treatment In the The planned tooth movement distance in the patient's medical record during the first visit; For patients with a history of orthodontic treatment In the The degree of local correction achievement in the medical record of the first visit; For patients with a history of orthodontic treatment The degree of orthodontic achievement; It is an exponential function with the natural constant as its base; For patients with a history of orthodontic treatment The degree of indirect response performance in correction.

[0073] The scheme determination module 104 is used to determine the invisible orthodontic treatment plan for the patient to be analyzed in the similar patient clusters in which the patient to be analyzed belongs, based on the referenceability of the corresponding samples and the characteristic deviation distance between each historical orthodontic patient and the patient to be analyzed.

[0074] After determining the cluster of similar patients to which the patient to be analyzed belongs based on the characteristic deviation distance between the patient and each historical patient, this invention further considers that the greater the reference value of the sample in the cluster, the broader the reference value of the corresponding data sample. This necessitates a greater focus on the comprehensive treatment response, which characterizes the treatment effects of each historical orthodontic treatment plan for each patient, when selecting a clear aligner treatment plan. Conversely, the lower the reference value of the sample, the worse the overall reference value of the cluster's data sample. Therefore, it is more appropriate to select a clear aligner treatment plan based on the historical orthodontic treatment plans of patients with more similar multi-dimensional characteristics. This comprehensive analysis allows the patient to achieve better orthodontic performance. Therefore, this embodiment of the invention determines the clear aligner treatment plan for the patient to be analyzed within the cluster of similar patients, based on the corresponding sample reference value and the characteristic deviation distance between each historical patient and the patient to be analyzed.

[0075] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining a dental invisible orthodontic treatment plan includes:

[0076] Based on the referenceability of the sample clusters of similar patients to which the orthodontic patient to be analyzed belongs, a first reference weight is determined. The first reference weight is then negatively correlated to determine a second reference weight. The sum of the first and second reference weights is 1. Historical orthodontic patients within the similar patient clusters to which the orthodontic patient to be analyzed belong are used as reference patients. Since a higher sample referenceability necessitates a greater emphasis on the overall treatment response of each reference patient, while a lower sample referenceability necessitates a greater emphasis on the characteristic deviation distance between each reference patient and the orthodontic patient to be analyzed, a first and second reference weight are determined based on the sample referenceability. The first reference weight represents the weight of the overall treatment response, and the second reference weight represents the weight of the characteristic deviation distance.

[0077] Therefore, this embodiment of the invention further determines the weighted response of each reference patient based on the product of the normalized value of the comprehensive treatment response of each reference patient and the first reference weight; and determines the weighted similarity of each reference patient based on the second reference weight and the feature deviation distance between each reference patient and the orthodontic patient to be analyzed. The process of obtaining the weighted similarity includes: performing a negative correlation mapping on the feature deviation distance between each reference patient and the orthodontic patient to be analyzed to determine the corresponding feature similarity; and determining the weighted similarity of each reference patient based on the product of the feature similarity and the second reference weight. It should be noted that, unless otherwise specified, the normalization method in this embodiment of the invention adopts the min-max normalization method, which will not be further elaborated here.

[0078] This invention employs a negative correlation mapping between the feature deviation distance between each reference patient and the orthodontic patient to be analyzed, thus limiting the obtained feature similarity to between 0 and 1. Therefore, when weighting the overall treatment response and feature similarity using the first and second reference weights respectively, the corresponding values ​​are all within the range of 0-1, avoiding the influence of dimensions on the calculation process. Further details are omitted here. By weighting the overall treatment response and feature similarity with weights of different magnitudes and correlations, the subsequent calculation of the fit optimization degree takes into account the emphasis on different calculation dimensions under the influence of sample referenceability. This aligns with the objective fact that a smaller feature deviation distance between the reference patient and the orthodontic patient to be analyzed corresponds to a larger overall treatment response, indicating a better fit of the historical orthodontic treatment plan for the patient to be analyzed. This makes the subsequent invisible orthodontic treatment plan selected based on the fit optimization degree more accurate.

[0079] Therefore, based on the correlation, the fit optimization degree of each reference patient is determined according to the sum of the weighted similarity and weighted responsiveness. Based on the fit optimization degree and the historical orthodontic treatment plan of each reference patient, the invisible orthodontic treatment plan of the patients to be analyzed is selected. Specifically, the historical orthodontic treatment plan of the reference patient with the highest fit optimization degree is used as the invisible orthodontic treatment plan of the patients to be analyzed.

[0080] In one specific implementation of this invention, the process of obtaining the adaptation optimization degree is expressed by the following formula: ; Reference patients for the orthodontic patients to be analyzed The degree of adaptation and optimization; The first reference weight is the sample reference value of the similar patient cluster to which the orthodontic patient to be analyzed belongs; Reference patients for the orthodontic patients to be analyzed The overall response rate of treatment; Reference patients for the orthodontic patients to be analyzed The weighted response; The second reference weights are assigned to the orthodontic patients to be analyzed. To analyze orthodontic patients and corresponding reference patients The characteristic deviation distance between them; To analyze orthodontic patients and corresponding reference patients Feature similarity between them; To analyze orthodontic patients and corresponding reference patients Weighted similarity between them.

[0081] In summary, an AI-powered invisible orthodontic treatment plan generation system first divides patients into similar patient clusters based on the similarity of their historical orthodontic treatments in terms of dental malformation symptom sequences and physiological feature data sequences. Then, it determines the overall treatment response based on the treatment response characteristics represented by the actual tooth movement distance deviation and complication frequency distribution of each historical orthodontic patient. Finally, it comprehensively determines the invisible orthodontic treatment plan for the patient under analysis by weighting the sample referenceability of the sample size within the similar patient cluster to which the patient to be analyzed belongs, combined with the feature deviation distance and the overall treatment response, thus resulting in a more accurate invisible orthodontic treatment plan.

[0082] This application also provides a computer device; please refer to [link / reference]. Figure 2 The illustration shows a schematic diagram of a computer device structure provided by an embodiment of the present invention. The computer device includes a memory 201, a processor 202, and a computer program 203 stored in the memory 201 and running on the processor 202. When the processor 202 executes the computer program 203, the computer device can execute any of the aforementioned AI generation systems for invisible orthodontic treatment plans.

[0083] This application also provides a computer program product that, when run on a computer device, enables the computer device to execute any of the aforementioned AI-generated invisible orthodontic treatment plans.

[0084] This application also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer device, the computer device can execute any of the aforementioned AI-generated invisible orthodontic treatment plans.

[0085] In the embodiments provided in this application, it should be understood that the computer device, computer program product and computer-readable storage medium provided are all used to execute the corresponding system provided above, and therefore the beneficial effects they can achieve can be referred to the beneficial effects of the system provided above, which will not be repeated here.

[0086] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0087] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. An AI-generated system for invisible orthodontic treatment plans, characterized in that, The system includes: The data acquisition and preprocessing module is used to obtain the physiological characteristic data sequence, dental malocclusion symptom sequence, actual tooth movement distance and all complications of each historical orthodontic patient from the medical database; wherein the historical orthodontic treatment plan of each historical orthodontic patient is the same in all the medical records of all visits. The first determination module is used to determine the corresponding feature deviation distance based on the similarity of the dental malocclusion symptom sequence and physiological characteristic data sequence between each historical orthodontic patient and other orthodontic patients; to perform cluster analysis based on the feature deviation distance between historical orthodontic patients to determine all similar patient clusters; and to determine the corresponding sample reference degree based on the number of historical orthodontic patients in each similar patient cluster. The process of obtaining the feature deviation distance includes: Each historical orthodontic patient was selected as the target orthodontic patient, and other historical orthodontic patients were selected as reference orthodontic patients. The similarity of dental malocclusion symptoms is determined based on the Jaccard coefficient between the dental malocclusion symptom sequence of the target orthodontic patient and the dental malocclusion symptom sequence of each reference orthodontic patient. The Euclidean distance between the physiological characteristic data sequence of the target orthodontic patient and the physiological characteristic data sequence of each reference orthodontic patient is normalized to obtain the corresponding physiological characteristic deviation. The feature deviation distance between the target orthodontic patient and each reference orthodontic patient is determined by multiplying the negative correlation mapping value of the similarity of the dental malocclusion symptoms with the deviation of the physiological features. The second determination module is used to determine the corresponding overall treatment response based on the planned deviation of the actual tooth movement distance and the distribution of complication frequency in the treatment records of each historical orthodontic patient throughout all visits. The process of obtaining the overall treatment response includes: The orthodontic achievement rate is determined based on the overall deviation between the actual tooth movement distance and the corresponding planned tooth movement distance in the treatment records of each patient throughout all visits. The indirect response performance of orthodontics was determined based on the frequency distribution of complications in the treatment records of each patient throughout all visits. The corresponding overall treatment response is determined by multiplying the orthodontic achievement rate with the indirect response performance of the orthodontic treatment. The treatment plan determination module is used to determine the invisible orthodontic treatment plan for the patient to be analyzed within the similar patient clusters in which the patient to be analyzed belongs, based on the referenceability of the corresponding samples and the characteristic deviation distance between each historical orthodontic patient and the patient to be analyzed. The process of obtaining the invisible orthodontic treatment plan includes: Based on the referenceability of samples in the similar patient clusters to which the orthodontic patient to be analyzed belongs, a first reference weight is determined; the first reference weight is negatively correlated to determine a second reference weight; the sum of the first reference weight and the second reference weight is 1; historical orthodontic patients in the similar patient clusters to which the orthodontic patient to be analyzed belongs are used as reference patients. The weighted response of each reference patient is determined by multiplying the normalized value of the overall treatment response of each reference patient with the first reference weight; the weighted similarity of each reference patient is determined by the second reference weight and the feature deviation distance between each reference patient and the orthodontic patient to be analyzed. The fit optimization degree of each reference patient is determined based on the sum of the weighted similarity and the weighted responsiveness. Based on the fit optimization and the historical orthodontic treatment plans of each reference patient, invisible orthodontic treatment plans for patients to be analyzed were selected.

2. The AI-generated invisible orthodontic treatment plan system according to claim 1, characterized in that, The process of obtaining the similar patient clusters includes: Based on the feature deviation distance, the K-means clustering algorithm was used to perform cluster analysis on all historical orthodontic patients to determine all similar patient clusters.

3. The AI-generated invisible orthodontic treatment plan system according to claim 1, characterized in that, The process of obtaining the referenceability of the sample includes: The number of patients with historical orthodontic treatment in each similar patient cluster is normalized to determine the reference value of the corresponding sample.

4. The AI-generated invisible orthodontic treatment plan system according to claim 1, characterized in that, The process of obtaining the orthodontic achievement rate includes: The difference between the actual tooth movement distance and the corresponding planned tooth movement distance in the treatment records of each historical orthodontic patient at each visit was normalized to determine the corresponding local orthodontic achievement; the corresponding orthodontic achievement was determined based on the mean of the local orthodontic achievement in the treatment records of each historical orthodontic patient across all visits.

5. The AI-generated invisible orthodontic treatment plan system according to claim 1, characterized in that, The process of obtaining the corrective indirect response performance includes: The frequency of complications for each historical orthodontic patient was arranged chronologically from all treatment records and then curve-fitted to determine the complication frequency curve. The mean of the incisional slope corresponding to all treatments on the complication frequency curve was negatively correlated to determine the trend value of complication changes. The overall complication severity was determined based on the mean frequency of complications in all treatment records of each historical orthodontic patient. The orthodontic indirect response performance of each patient with a history of orthodontic treatment is determined by multiplying the negative correlation mapping value of the overall performance of the complications with the trend value of the complications.

6. The AI-generated invisible orthodontic treatment plan system according to claim 1, characterized in that, The process of obtaining the weighted similarity includes: The feature deviation distance between each reference patient and the orthodontic patient to be analyzed is negatively correlated to determine the corresponding feature similarity; the weighted similarity of each reference patient is determined based on the product of the feature similarity and the second reference weight.

7. The AI-generated invisible orthodontic treatment plan system according to claim 1, characterized in that, The process of selecting invisible orthodontic treatment plans for patients to be analyzed based on the fit optimization and the historical orthodontic treatment plans of each reference patient includes: The historical orthodontic treatment plan of the reference patient with the highest fit and optimization degree will be used as the invisible orthodontic treatment plan for the patient to be analyzed.

Images