Intelligent processing method for dental film data
By constructing a decision tree and utilizing multiple feature values and orthodontic correction coefficients from dental radiographs, subtle changes in dimensions can be identified, thus solving the problem of inaccurate classification of dental radiographs in decision tree models and improving the accuracy of dental radiograph data processing.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-25
- Publication Date
- 2026-04-03
AI Technical Summary
In existing technologies, the varying sensitivity of different tooth feature dimensions to smoothness issues in decision tree models results in significant fluctuations in model predictions, making accurate classification impossible.
By acquiring multiple feature values from dental X-ray data, a decision tree is constructed. Orthodontic correction coefficients and smoothing sensitivity analysis are used to identify subtle change dimensions. Based on the subtle change indicators, a decision tree is constructed to improve classification accuracy.
This improved the accuracy of dental X-ray data in decision tree classification, reduced the impact of minor changes, and achieved more accurate dental X-ray data processing.
Smart Images
Figure CN121786632A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of dental radiograph data processing technology, and more specifically to an intelligent processing method for dental radiograph data. Background Technology
[0002] Thanks to technological advancements and increased public awareness of oral health, the orthodontic industry has flourished in recent years. By combining computer-aided 3D diagnosis, personalized design, and digital prototyping technologies, medical professionals can provide more precise and efficient orthodontic solutions. In this industry, the processing of dental X-ray data has become particularly crucial, and the introduction of artificial intelligence technology has made intelligent processing of this data possible, significantly improving the efficiency and accuracy of orthodontic treatment.
[0003] Because dental X-ray data is complex data with non-linear relationships, existing technologies often use decision tree algorithms to classify it and determine whether the dental X-ray data is normal. However, dental X-ray data contains multiple dimensions of tooth features, and different tooth feature dimensions may have different sensitivities to smoothness issues in the decision tree model. Some tooth feature dimensions may be more easily affected by small changes, causing large fluctuations in the model's predictions, while other tooth feature dimensions may be less sensitive to smoothness issues, thus leading to inaccurate classification of the dental X-ray data. Summary of the Invention
[0004] To address the issue that different tooth feature dimensions may exhibit varying sensitivities to smoothness in decision tree models—some dimensions being more susceptible to subtle changes leading to significant fluctuations in predictions, while others may be less sensitive to smoothness, resulting in inaccurate classification of dental radiograph data—this invention aims to provide an intelligent processing method for dental radiograph data. The specific technical solution adopted is as follows:
[0005] A method for intelligent processing of dental radiograph data, the method comprising:
[0006] Acquire all dental radiographs for each patient; each radiograph includes feature values for the dental characteristic dimensions of occlusion, axial tilt, torque angle, crowding, and Spee curvature.
[0007] A decision tree is constructed based on the occlusal relationship of all dental radiographs to classify all the radiographs and obtain all first malocclusion categories. Orthodontic correction coefficients are obtained for each patient's radiographs based on tooth position distribution and the number of missing teeth. One first malocclusion category is randomly selected as a reference first malocclusion category. The reference first malocclusion category is then classified according to the orthodontic correction coefficients to obtain all second malocclusion categories. In addition to occlusal relationship, one dental feature dimension is randomly selected as a reference feature dimension. The smoothness sensitivity of the reference feature dimension is obtained based on the feature value changes of all radiographs in each second malocclusion category across adjacent treatment stages and the orthodontic correction coefficients. Based on the smoothness sensitivity, minute change dimensions within the dental feature dimension are obtained. One minute change dimension is randomly selected as a reference micro-change dimension. The minute change index for each consecutive treatment stage interval is obtained based on the feature value changes under the reference micro-change dimension across all consecutive treatment stage intervals for each patient.
[0008] Based on the distribution of the micro-change indicators, obtain the micro-change dental radiograph data interval for each micro-change dimension; construct a decision tree for the dental radiograph data based on all micro-change dental radiograph data intervals.
[0009] Furthermore, the method for obtaining the orthodontic correction coefficient includes:
[0010] The orthodontic correction coefficient is obtained according to the orthodontic correction coefficient calculation formula, which is as follows:
[0011]
[0012] In the formula, C i denoted as the orthodontic correction coefficient for the i-th patient; N represents the actual number of teeth in the patient; 32 represents the ideal number of teeth in the patient; α represents the initial angle of the axial tilt of the n-th tooth in the patient; α′ represents the expected angle of the axial tilt of the n-th tooth in the patient; β represents the initial angle of the torque angle of the n-th tooth in the patient; β′ represents the expected angle of the torque angle of the n-th tooth in the patient; γ represents the crowding of the n-th tooth in the patient; || represents the absolute value function; exp() represents the exponential function with the natural constant as the base.
[0013] Furthermore, a decision tree is constructed based on the occlusal relationships of all dental radiographs to classify all dental radiographs and obtain all first-order malocclusion categories, including:
[0014] The occlusal relationships include neutral malocclusion, distal malocclusion, and mesial malocclusion; the neutral malocclusion, the distal malocclusion, and the mesial malocclusion are each classified as a first type of malocclusion.
[0015] Furthermore, the method for obtaining the second deformity category includes:
[0016] Dental radiographs of all patients who fall under the first reference malformation category are used as the first dental radiograph data;
[0017] The orthodontic correction coefficients of the first dental radiograph data are normalized to obtain the basic correction coefficients for each first dental radiograph data; all first dental radiograph data are divided into four categories according to the quartile algorithm, and each category is used as the second malocclusion category.
[0018] Furthermore, the method for obtaining the smoothness sensitivity includes:
[0019] The smoothing sensitivity is obtained according to the smoothing sensitivity calculation formula, which is shown below:
[0020]
[0021] In the formula, D represents the smoothness sensitivity of the reference feature dimension; J represents the number of patients in each second malformation category; M represents the number of treatment stages required for each patient in the second malformation category under the reference feature dimension, which can be obtained by existing technology; V m+2 V represents the feature value of each patient in the second malformation category at the beginning of the (m+2)th treatment phase under the reference feature dimension; m+1 V represents the feature value of each patient in the second malformation category at the beginning of the (m+1)th treatment phase under the reference feature dimension; m H represents the feature value of each patient in the second malformation category at the beginning of the m-th treatment phase under the reference feature dimension; m+1 This represents the treatment duration for each patient in the second malformation category at the (m+1)th treatment stage under the reference feature dimension, which can be obtained using existing technology; H m σ represents the treatment duration of each patient in the second malformation category at the m-th treatment stage under the reference feature dimension; Ci represents the standard deviation of the orthodontic correction coefficients for all dental radiographs of all patients in the second malocclusion category; | represents the absolute value function.
[0022] Furthermore, the method for obtaining the dimension of minute changes includes:
[0023] Each tooth feature dimension whose smoothness sensitivity is greater than a preset first threshold is taken as a tiny change dimension.
[0024] Furthermore, the method for obtaining the minute change index includes:
[0025] When the number of treatment stages contained in each consecutive treatment stage interval is 1, under the reference micro-variation dimension, the feature value at the beginning of the treatment stage is used as the lower limit of the feature value interval, and the feature value at the end of the treatment stage is used as the upper limit of the feature value interval to obtain the feature value interval.
[0026] When the number of treatment stages contained in each consecutive treatment stage interval is greater than 1, under the reference micro-variation dimension, the initial feature value of the first treatment stage interval in each consecutive treatment stage is taken as the lower limit of the feature value interval, and the feature value at the end of the last treatment stage is taken as the upper limit of the feature value interval, thus obtaining the feature value interval.
[0027] The minute change index is obtained according to the formula for calculating the minute change index, which is as follows:
[0028]
[0029] In the formula, K represents the small change index of the characteristic value interval; L represents the number of treatment stages contained in the continuous treatment stage interval corresponding to the characteristic value interval; F′ represents the upper limit of the characteristic value interval; F represents the lower limit of the characteristic value interval; exp() represents the exponential function with the natural constant as the base.
[0030] Furthermore, the method for obtaining the range of minutely changing dental radiograph data includes:
[0031] The eigenvalue interval corresponding to the maximum minute change index of each patient under the reference minute change dimension is obtained as the first eigenvalue interval; the intersection of the first eigenvalue intervals of all patients is taken as the minute change dental radiograph data interval of the reference minute change dimension.
[0032] An intelligent processing system for dental radiograph data is provided. The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the intelligent processing method for dental radiograph data described above.
[0033] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the intelligent processing method for dental radiograph data as described above.
[0034] The present invention has the following beneficial effects:
[0035] This invention acquires all dental radiographs for each patient. Since medical professionals describe tooth morphology and function based on various characteristics of the patient's teeth, each radiograph includes tooth feature dimensions such as occlusal relationship, axial tilt angle, torque angle, crowding degree, and Spee curvature. Because the radiograph data is non-linear, a decision tree algorithm is used for processing. In reality, medical professionals classify malocclusion into neutral, distal, and mesial malocclusions; therefore, all radiographs are classified according to occlusal relationship to obtain all first-order malocclusion categories. Since the higher the degree of dental malocclusion, the longer the treatment time required, each patient's radiograph data... The orthodontic correction coefficients were analyzed and used as the classification basis for the second layer of the decision tree, obtaining all nodes in the second layer of the decision tree. Since axial tilt, torque angle, crowding, and Spee curvature can reflect the dental characteristics of different patients, they were used as the classification basis for other layers of the decision tree. Because small changes may cause similar dental radiograph data to be classified into different nodes, the small change indicators of the feature value intervals corresponding to the continuous treatment stages were analyzed. Based on the distribution of the small change indicators, the small change dental radiograph data intervals for each small change dimension were obtained. A decision tree was constructed based on all the small change dental radiograph data intervals. This invention makes different dental feature dimensions less susceptible to the influence of small changes in dental radiograph data during decision tree classification, improving the classification accuracy of the decision tree and thus making the intelligent processing of dental radiograph data more accurate. Attached Figure Description
[0036] 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.
[0037] Figure 1 This is a flowchart of an intelligent processing method for dental radiograph data provided in one embodiment of the present invention. Detailed Implementation
[0038] 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 intelligent processing method for dental X-ray data proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.
[0039] 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.
[0040] The following description, in conjunction with the accompanying drawings, details the specific scheme of the intelligent processing method for dental X-ray data provided by the present invention.
[0041] Please see Figure 1 This illustrates an intelligent processing method for dental radiograph data according to an embodiment of the present invention, the method comprising:
[0042] Step S1: Obtain all dental radiographs for each patient; each radiograph includes feature values for the dental characteristic dimensions such as occlusal relationship, axial tilt, torque angle, crowding, and Spee curvature.
[0043] This invention is primarily applied to intelligent classification scenarios for dental X-ray data. Therefore, the first step is to acquire all dental X-ray data for each patient. Since medical professionals describe tooth morphology and function based on various features of a patient's teeth, such as occlusion, axial tilt, torque angle, crowding, and Spee curvature, this invention requires collecting feature values for each patient's occlusion, axial tilt, torque angle, crowding, and Spee curvature dimensions from the acquired dental X-ray data.
[0044] The aforementioned occlusion relationship, axial tilt angle, torque angle, crowding degree, and Spee curvature are all technical means well known to those skilled in the art and can be directly obtained through the systems of hospitals or medical institutions, and will not be elaborated here.
[0045] In one embodiment of the present invention, the feature values of all tooth feature dimensions obtained above for each patient at each preset sampling time are combined into a vector, and this vector is used as the dental radiograph data for each patient. The preset sampling time is set as the beginning and end of each treatment phase, and the treatment phase can be set by different medical personnel, which is not limited here.
[0046] Step S2: Construct a decision tree for all dental radiographs based on the occlusal relationship, classify all dental radiographs to obtain all first malocclusion categories; obtain the orthodontic correction coefficient for each patient's dental radiographs based on the tooth position distribution and the number of missing teeth; randomly select one first malocclusion category as a reference first malocclusion category; classify the reference first malocclusion category according to the orthodontic correction coefficient to obtain all second malocclusion categories; randomly select a tooth feature dimension as a reference feature dimension, excluding the occlusal relationship; obtain the smoothness sensitivity of the reference feature dimension based on the feature value changes and orthodontic correction coefficients of all dental radiographs in each second malocclusion category across adjacent treatment stages; obtain the minute change dimension in the tooth feature dimension based on the smoothness sensitivity; randomly select a minute change dimension as a reference micro-change dimension; obtain the minute change index of the feature value interval corresponding to each consecutive treatment stage interval based on the feature value changes under the reference micro-change dimension across all consecutive treatment stage intervals for each patient.
[0047] Because dental radiograph data contains feature values from multiple dimensions of tooth characteristics, the data is quite complex, and different radiographs exhibit non-linear relationships. Therefore, a decision tree algorithm is used to process all dental radiograph data. In practice, the Angle classification system is commonly used to classify malocclusions. Medical personnel can classify malocclusions into neutral, distal, and mesial malocclusions based on the relative positional relationship between the maxillary first permanent molar and mandibular teeth, i.e., the occlusal relationship. The dental radiograph data is classified according to these three categories, with each category serving as the primary malocclusion category. The Angle classification system is a well-known technique in the field and will not be elaborated upon here.
[0048] Based on the above process, three nodes are obtained in the first layer of the decision tree, dividing all dental radiograph data into three categories. Each node is then analyzed separately, and subsequent classification is performed on each node.
[0049] Even with similar occlusal relationships, patients may still exhibit varying degrees of dental malocclusion. For instance, the axial tilt, torque angle, and crowding of teeth can differ significantly among patients. The higher the degree of dental malocclusion, the longer the treatment time required. Furthermore, the number of missing teeth may also vary among patients; a larger number of missing teeth may result in a longer potential healing time. Therefore, in this embodiment of the invention, orthodontic correction coefficients are obtained for each patient's dental radiographs based on their tooth position distribution and the number of missing teeth. These orthodontic correction coefficients are then used as the classification criteria for the second layer of the decision tree.
[0050] Preferably, in this embodiment of the invention, the method for obtaining the orthodontic correction coefficient includes:
[0051] The orthodontic correction coefficient is obtained according to the formula shown below:
[0052]
[0053] In the formula, C i denoted as the orthodontic correction coefficient for each dental radiograph of the i-th patient; N represents the actual number of teeth in the patient; 32 represents the ideal number of teeth in the patient; α represents the initial angle of the axial tilt of the n-th tooth in the patient; α′ represents the expected angle of the axial tilt of the n-th tooth in the patient; β represents the initial angle of the torque angle of the n-th tooth in the patient; β′ represents the expected angle of the torque angle of the n-th tooth in the patient; γ represents the crowding degree of the n-th tooth in the patient, which can be directly obtained by existing technology; || represents the absolute value function; exp() represents the exponential function with the natural constant as the base.
[0054] In the orthodontic correction factor calculation formula, the more missing teeth a patient has (|32-N|), the longer the treatment time will be required. The larger the value, the greater the orthodontic correction coefficient for each dental radiograph of the i-th patient; the greater the difference between the actual and expected characteristic values of the torque angle and axial tilt angle of each tooth in each dental radiograph of the i-th patient, the longer the adjustment time required for that radiograph, and the greater the orthodontic correction coefficient for that radiograph; the greater the crowding of the n-th tooth of the patient, the greater the degree of dental malocclusion in each dental radiograph of the patient, and the greater the orthodontic correction coefficient for each dental radiograph of the i-th patient.
[0055] Preferably, in one embodiment of the present invention, after obtaining the orthodontic correction coefficient for each dental radiograph of each patient, the dental radiographs of all patients that meet the reference first malocclusion category are taken as the first dental radiographs; the orthodontic correction coefficients of the first dental radiographs are normalized to obtain the basic correction coefficient for each first dental radiograph; and all the first dental radiographs are divided into four categories according to the quartile algorithm, with each category being taken as the second malocclusion category.
[0056] This completes the process of obtaining all nodes in the second layer of the decision tree.
[0057] Each node in the second layer is analyzed separately. Since axial tilt, torque angle, crowding, and Spee curvature can reflect the tooth characteristics of different patients, these four parameters are used as the classification criteria in the subsequent decision tree classification process. The analysis method for each tooth feature dimension used as the classification criterion is the same. Therefore, in the following embodiment, axial tilt is used as the classification criterion for the third layer of the decision tree for subsequent analysis.
[0058] It should be noted that in one embodiment of the present invention, the decision tree mainly relies on the information gain ratio when selecting the classification criteria. The calculation method of the information gain ratio is a well-known prior art and will not be described in detail here.
[0059] Because decision tree algorithms are overly sensitive to small changes in input data, even minor variations can lead to significant output changes. Therefore, it's necessary to determine which dental feature dimensions are prone to minute changes. If the number of treatment stages and the total treatment time required for each patient to adjust the radiograph data for a particular dental feature dimension are inconsistent, then the feature value of that dental feature dimension is considered volatile during treatment. Furthermore, the more discrete the orthodontic correction coefficients corresponding to each patient's radiograph data, the more volatile the feature value of that dental feature dimension is during treatment. Therefore, in this embodiment of the invention, based on the feature value changes and orthodontic correction coefficients of all radiograph data in each second malocclusion category across adjacent treatment stages on the reference feature dimension, the smoothness sensitivity of the reference feature dimension is obtained, and thus, the dimension of minute changes in the dental feature dimension is determined.
[0060] Preferably, in one embodiment of the present invention, the method for obtaining the smoothness sensitivity includes:
[0061] The smoothing sensitivity is obtained according to the smoothing sensitivity calculation formula, which is shown below:
[0062]
[0063] In the formula, D represents the smoothness sensitivity of the reference feature dimension; J represents the number of patients in each second malformation category; M represents the number of treatment stages required for each patient in the second malformation category under the reference feature dimension, which can be obtained by existing technology; V m+2 V represents the feature value of the dental radiograph data at the beginning of the (m+2)th treatment stage for each patient in the second malformation category, under the reference feature dimension; m+1 V represents the feature value of the dental radiograph data at the beginning of the (m+1)th treatment stage for each patient in the second malformation category, under the reference feature dimension; m H represents the feature value of the dental radiograph data at the beginning of the m-th treatment stage for each patient in the second malformation category, under the reference feature dimension; m+1 This represents the treatment duration for each patient in the second malformation category at the (m+1)th treatment stage under the reference feature dimension, which can be obtained using existing technology; H m This represents the treatment duration of each patient in the second malformation category at the m-th treatment stage under the reference feature dimension; represents the standard deviation of the orthodontic correction coefficients for each patient's dental radiographs in the second malocclusion category; | represents the absolute value function.
[0064] In the formula for calculating smoothness sensitivity, the smaller the difference in feature values of dental radiographs between any two adjacent treatment stages, that is, the more stable the changes in feature values of dental radiographs during treatment stages, the better. The closer it is to 1, the better.
[0065] The smaller the value, the more similar the treatment time between two consecutive treatment phases.
[0066] The closer it is to 1, The smaller; for all patients in each second malformation category The smaller the sum, the less likely the dental radiograph data under the reference feature dimension in each second malocclusion category is to change, and the lower the smoothness sensitivity of the reference feature dimension. Conversely, the smaller the standard deviation of the orthodontic correction coefficients for each patient's dental radiograph data in the second malocclusion category, the more compact the data across all radiographs within that category. The standard deviation of the orthodontic correction coefficients is then used as the basis for determining the optimal data set.
[0067] The weighting coefficients are such that the reference feature dimension is more likely to change during treatment, and the smoothness sensitivity of the reference feature dimension is greater.
[0068] Preferably, in one embodiment of the present invention, the method for obtaining the dimension of minute changes includes:
[0069] Each tooth feature dimension with a smoothing sensitivity greater than a preset first threshold is considered a minute change dimension. In one embodiment of the present invention, the preset first threshold is set to 0.8. It should be noted that in other embodiments of the present invention, the preset first threshold can be set arbitrarily, and is not limited here.
[0070] Since minute changes can cause dental X-ray data with similar features to be assigned to different nodes, thus affecting the predictive performance of the decision tree model, splitting this interval during the construction of the decision tree may cause the decision tree model to become unstable within this interval. Therefore, in this embodiment of the invention, the minute change index of the feature value interval corresponding to each consecutive treatment stage interval is obtained based on the feature value changes under the reference minute change dimension in all consecutive treatment stage intervals for each patient.
[0071] Preferably, in a personal embodiment of the present invention, the method for obtaining the minute change index includes:
[0072] When the number of treatment stages contained in each consecutive treatment stage interval is 1, under the reference micro-variation dimension, the feature value of the dental radiograph data at the beginning of the treatment stage is used as the lower limit of the feature value interval, and the feature value of the dental radiograph data at the end of the treatment stage is used as the upper limit of the feature value interval to obtain the feature value interval.
[0073] When the number of treatment stages contained in each consecutive treatment stage interval is greater than 1, under the reference micro-variation dimension, the initial feature value of the first treatment stage interval in each consecutive treatment stage is taken as the lower limit of the feature value interval, and the feature value at the end of the last treatment stage is taken as the upper limit of the feature value interval, thus obtaining the feature value interval.
[0074] The minute change index is obtained based on the formula for calculating the minute change index, which is shown below:
[0075]
[0076] In the formula, K represents the small change index of the characteristic value interval; L represents the number of treatment stages contained in the continuous treatment stage interval corresponding to the characteristic value interval; F′ represents the upper limit of the characteristic value interval; F represents the lower limit of the characteristic value interval; exp() represents the exponential function with the natural constant as the base.
[0077] In the formula for calculating the small change index, the rate of change of the eigenvalue is defined as the reference small change dimension. The smaller the value, the smaller the trend of characteristic value change in each continuous treatment stage interval under the reference micro-variation dimension, and the larger the micro-change index of the characteristic value interval. The larger the number of treatment stages contained in the continuous treatment stage interval corresponding to the characteristic value interval, the larger the time span required for treatment, and the larger the micro-change index of the characteristic value interval.
[0078] Step S3: Based on the distribution of the minute change indicators, obtain the minute change dental radiograph data intervals for each minute change dimension; construct a decision tree for the dental radiograph data based on all minute change dental radiograph data intervals.
[0079] Preferably, in one embodiment of the present invention, the method for obtaining the range of dental radiograph data with minute changes includes:
[0080] The eigenvalue interval corresponding to the maximum minute change index of each patient under the reference minute change dimension is obtained as the first eigenvalue interval; the intersection of the first eigenvalue intervals of all patients is taken as the minute change dental radiograph data interval of the reference minute change dimension.
[0081] The small variation range of dental radiograph data corresponding to each tooth feature dimension is added to the classification criteria of the decision tree to obtain the decision tree node corresponding to each tooth dimension, thus completing the construction of the decision tree. In one embodiment of the present invention, the C4.5 decision tree model is used for operation. It should be noted that the C4.5 decision tree algorithm is a well-known technique in the art and will not be limited or described in detail here.
[0082] This completes the intelligent processing of the dental X-ray data.
[0083] In summary, all dental radiographs for each patient were acquired. Each radiograph included feature values for tooth characteristics such as occlusion, axial tilt, torque angle, crowding, and Spee curvature. A decision tree was constructed based on the occlusion of all radiographs to classify them and obtain all first-order malocclusion categories. Orthodontic correction coefficients were obtained for each patient based on tooth position distribution and the number of missing teeth. One first-order malocclusion category was randomly selected as a reference first-order malocclusion category. The reference first-order malocclusion category was then classified based on the orthodontic correction coefficients to obtain all second-order malocclusion categories. In addition to occlusion, one tooth characteristic dimension was randomly selected as a reference feature dimension. The process involves: determining the smoothness sensitivity of the reference feature dimension based on the changes in feature values and orthodontic correction coefficients across all dental radiographs in each second malocclusion category during adjacent treatment phases; determining the minute change dimension within the dental feature dimension based on the smoothness sensitivity; selecting one minute change dimension as the reference micro-change dimension; determining the minute change index for the feature value interval corresponding to each consecutive treatment phase based on the changes in feature values under the reference micro-change dimension across all consecutive treatment phase intervals for each patient; determining the minute change dental radiograph data interval for each minute change dimension based on the distribution of the minute change index; and constructing a decision tree for the dental radiograph data based on all minute change dental radiograph data intervals.
[0084] One embodiment of the present invention provides an intelligent processing system for dental radiograph data. The system includes a memory, a processor, and a computer program. The memory is used to store the corresponding computer program, and the processor is used to run the corresponding computer program. When the computer program runs in the processor, it can implement the methods described in steps S1-S4.
[0085] This invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the intelligent processing method for dental radiograph data described above.
[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. A method for intelligent processing of dental radiograph data, characterized in that, The method includes: Acquire all dental radiographs for each patient; each radiograph includes feature values for the dental characteristic dimensions of occlusion, axial tilt, torque angle, crowding, and Spee curvature. A decision tree is constructed based on the occlusal relationship of all dental radiographs to classify all the radiographs and obtain all first malocclusion categories. Orthodontic correction coefficients are obtained for each patient's radiographs based on tooth position distribution and the number of missing teeth. One first malocclusion category is randomly selected as a reference first malocclusion category. The reference first malocclusion category is then classified according to the orthodontic correction coefficients to obtain all second malocclusion categories. In addition to occlusal relationship, one dental feature dimension is randomly selected as a reference feature dimension. The smoothness sensitivity of the reference feature dimension is obtained based on the feature value changes of all radiographs in each second malocclusion category across adjacent treatment stages and the orthodontic correction coefficients. Based on the smoothness sensitivity, minute change dimensions within the dental feature dimension are obtained. One minute change dimension is randomly selected as a reference micro-change dimension. The minute change index for each consecutive treatment stage interval is obtained based on the feature value changes under the reference micro-change dimension across all consecutive treatment stage intervals for each patient. Based on the distribution of the micro-change indicators, obtain the micro-change dental radiograph data interval for each micro-change dimension; construct a decision tree for the dental radiograph data based on all micro-change dental radiograph data intervals.
2. The intelligent processing method for dental radiograph data according to claim 1, characterized in that, The method for obtaining the orthodontic correction coefficient includes: The orthodontic correction coefficient is obtained according to the orthodontic correction coefficient calculation formula, which is as follows: In the formula, C i denoted as the orthodontic correction coefficient for the i-th patient; N represents the actual number of teeth in the patient; 32 represents the ideal number of teeth in the patient; α represents the initial angle of the axial tilt of the n-th tooth in the patient; α′ represents the expected angle of the axial tilt of the n-th tooth in the patient; β represents the initial angle of the torque angle of the n-th tooth in the patient; β′ represents the expected angle of the torque angle of the n-th tooth in the patient; γ represents the crowding of the n-th tooth in the patient; || represents the absolute value function; exp() represents the exponential function with the natural constant as the base.
3. The intelligent processing method for dental radiograph data according to claim 1, characterized in that, Based on the occlusal relationships of all dental radiographs, a decision tree is constructed for all dental radiographs to classify them and obtain all first-order malocclusion categories, including: The occlusal relationships include neutral malocclusion, distal malocclusion, and mesial malocclusion; the neutral malocclusion, the distal malocclusion, and the mesial malocclusion are each classified as a first type of malocclusion.
4. The intelligent processing method for dental radiograph data according to claim 1, characterized in that, The method for obtaining the second deformity category includes: Dental radiographs of all patients who fall under the first reference malformation category are used as the first dental radiograph data; The orthodontic correction coefficients of the first dental radiograph data are normalized to obtain the basic correction coefficients for each first dental radiograph data; all first dental radiograph data are divided into four categories according to the quartile algorithm, and each category is used as the second malocclusion category.
5. The intelligent processing method for dental radiograph data according to claim 4, characterized in that, The method for obtaining the smoothness sensitivity includes: The smoothing sensitivity is obtained according to the smoothing sensitivity calculation formula, which is shown below: In the formula, D represents the smoothness sensitivity of the reference feature dimension; J represents the number of patients in each second malformation category; M represents the number of treatment stages required for each patient in the second malformation category under the reference feature dimension, which can be obtained by existing technology; V m+2 V represents the feature value of each patient in the second malformation category at the beginning of the (m+2)th treatment phase under the reference feature dimension; m+1 V represents the feature value of each patient in the second malformation category at the beginning of the (m+1)th treatment phase under the reference feature dimension; m H represents the feature value of each patient in the second malformation category at the beginning of the m-th treatment phase under the reference feature dimension; m+1 This represents the treatment duration for each patient in the second malformation category at the (m+1)th treatment stage under the reference feature dimension, which can be obtained using existing technology; H m This represents the treatment duration of each patient in the second malformation category at the m-th treatment stage under the reference feature dimension; represents the standard deviation of the orthodontic correction coefficients for all dental radiographs of all patients in the second malocclusion category; || represents the absolute value function.
6. The intelligent processing method for dental radiograph data according to claim 1, characterized in that, The methods for obtaining the dimension of the minute change include: Each tooth feature dimension whose smoothness sensitivity is greater than a preset first threshold is taken as a tiny change dimension.
7. The intelligent processing method for dental radiograph data according to claim 1, characterized in that, The method for obtaining the minute change index includes: When the number of treatment stages contained in each consecutive treatment stage interval is 1, under the reference micro-variation dimension, the feature value at the beginning of the treatment stage is used as the lower limit of the feature value interval, and the feature value at the end of the treatment stage is used as the upper limit of the feature value interval to obtain the feature value interval. When the number of treatment stages contained in each consecutive treatment stage interval is greater than 1, under the reference micro-variation dimension, the initial feature value of the first treatment stage interval in each consecutive treatment stage is taken as the lower limit of the feature value interval, and the feature value at the end of the last treatment stage is taken as the upper limit of the feature value interval, thus obtaining the feature value interval. The minute change index is obtained according to the calculation formula for the minute change index, which is as follows: In the formula, K represents the small change index of the characteristic value interval; L represents the number of treatment stages contained in the continuous treatment stage interval corresponding to the characteristic value interval; F′ represents the upper limit of the characteristic value interval; F represents the lower limit of the characteristic value interval; exp() represents the exponential function with the natural constant as the base.
8. The intelligent processing method for dental radiograph data according to claim 1, characterized in that, The method for obtaining the range of dental radiograph data with minute changes includes: The eigenvalue interval corresponding to the maximum minute change index of each patient under the reference minute change dimension is obtained as the first eigenvalue interval; the intersection of the first eigenvalue intervals of all patients is taken as the minute change dental radiograph data interval of the reference minute change dimension.
9. An intelligent processing system for dental radiograph data, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent processing method for dental radiograph data as described in any one of claims 1 to 8.
10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent processing method for dental radiograph data as described in any one of claims 1 to 7.