A method and device for abnormal classification of dental data

Abstract

The present invention discloses a method and device for abnormal classification of tooth data, the method comprising: obtaining a tooth point cloud data set; the tooth point cloud data set comprises tooth point cloud data; preprocessing the tooth point cloud data set to obtain a preprocessed data set; using a trained abnormal recognition model to perform recognition processing on the preprocessed data set to obtain abnormal classification result information; the values ​​of the abnormal classification result information include normal, missing teeth, deformity, and dislocation. The present invention realizes efficient and accurate classification of tooth abnormalities based on tooth data.

Description

Technical Field

[0001] The present invention relates to the field of medical information processing, and in particular to a method and device for abnormal classification of tooth data. Background Art

[0002] Human teeth have their own distribution rules. They are distributed along the dental arch line. Various types of teeth are arranged from the center to both sides in a fixed order, but not all people's teeth can be arranged completely according to this rule. Many people have various problems such as deformed, misplaced, missing, and damaged teeth due to different reasons. This leads to various abnormalities in the teeth in medical images, which makes it very difficult to classify teeth in medical images.

[0003] At present, the identification of dental abnormalities mainly relies on doctors' experience, which has problems such as lack of objective standards and easy identification errors. How to achieve efficient and accurate abnormality classification based on dental data is an urgent problem that needs to be solved. Summary of the invention

[0004] In order to solve the problem of how to realize efficient and accurate classification of tooth abnormalities based on tooth data, the present invention discloses a method and device for classifying tooth abnormalities.

[0005] In a first aspect of an embodiment of the present application, a method for classifying abnormalities in tooth data is disclosed, comprising:

[0006] S1, obtaining a tooth point cloud data set; the tooth point cloud data set includes tooth point cloud data;

[0007] S2, preprocessing the tooth point cloud dataset to obtain a preprocessed dataset;

[0008] S3, using the trained abnormality recognition model to perform recognition processing on the preprocessed data set to obtain abnormality classification result information; the values ​​of the abnormality classification result information include normal, missing teeth, deformity, and dislocation.

[0009] The preprocessing of the tooth point cloud data set to obtain a preprocessed data set includes:

[0010] S21, performing noise reduction processing on the tooth point cloud dataset to obtain a noise reduction dataset;

[0011] S22, performing data cleaning processing on the noise reduction data set to obtain a first data set;

[0012] S23, performing point cloud registration processing on the first data set to obtain a preprocessed data set.

[0013] The anomaly recognition model includes: a feature fusion sub-model, a first encoding sub-model, a second encoding sub-model, and a third encoding sub-model;

[0014] The first coding sub-model is used to perform dislocation coding processing on the input tooth point cloud data to obtain a dislocation coding signal;

[0015] The second coding sub-model is used to perform deformity coding processing on the input tooth point cloud data to obtain a deformity coding signal;

[0016] The third coding sub-model is used to perform missing tooth coding processing on the input tooth point cloud data to obtain a missing tooth coding signal;

[0017] The feature fusion sub-model is connected to the first coding sub-model, the second coding sub-model and the third coding sub-model respectively, and is used to fuse the input misaligned coding signals, deformed coding signals and missing tooth coding signals to obtain abnormal classification result information.

[0018] The feature fusion sub-model includes a first input module, a first self-attention module, a first multi-head attention module, a first addition module, a first fusion output module, a second input module, a second self-attention module, a second multi-head attention module, and a second addition module;

[0019] The first fusion output module is used to output abnormal classification result information;

[0020] The first input module includes an input port, a first output port, and a second output port, and is used to receive the encoding information output by the first encoding sub-model, and output the encoding information through the first output port and the second output port respectively; the first output port of the first input module is connected to the input port of the first self-attention module; the second output port of the first input module is connected to the second input port of the first addition module;

[0021] The first self-attention module includes an input port, a first output port, a second output port, and a third output port; the first output port is used to output the Q variable after the encoding information output by the first encoding sub-model is self-attention encoded; the second output port is used to output the K variable after the encoding information output by the first encoding sub-model is self-attention encoded; the third output port is used to output the V variable after the encoding information output by the first encoding sub-model is self-attention encoded; the first output port of the first self-attention module is connected to the first input port of the first multi-head attention module; the second output port of the first self-attention module is connected to the second input port of the second multi-head attention module; the third output port of the first self-attention module is connected to the third input port of the second multi-head attention module;

[0022] The first multi-head attention module comprises a first input port, a second input port, a third input port and an output port; the output port of the first multi-head attention module is connected to the first input port of the first addition module;

[0023] The first addition module is used to sum the input variable of the first input port and the input variable of the second input port to obtain a sum variable, and send the sum variable to the first fusion output module;

[0024] The second input module includes an input port, a first output port, and a second output port, and is used to receive the encoding information output by the second encoding sub-model, and output the encoding information through the first output port and the second output port respectively; the first output port of the second input module is connected to the input port of the second self-attention module; the second output port of the second input module is connected to the second input port of the second addition module;

[0025] The second self-attention module includes an input port, a first output port, a second output port, and a third output port; the first output port is used to output the Q variable after the encoding information output by the third encoding sub-model is self-attention encoded; the second output port is used to output the K variable after the encoding information output by the third encoding sub-model is self-attention encoded; the third output port is used to output the V variable after the encoding information output by the third encoding sub-model is self-attention encoded; the first output port of the second self-attention module is connected to the first input port of the second multi-head attention module; the second output port of the second self-attention module is connected to the second input port of the second multi-head attention module; the third output port of the second self-attention module is connected to the third input port of the second multi-head attention module;

[0026] The second multi-head attention module comprises a first input port, a second input port, a third input port and an output port; the output port of the second multi-head attention module is connected to the first input port of the second addition module;

[0027] The second addition module is used to sum the input variable of the first input port and the input variable of the second input port to obtain a sum variable, and send the sum variable to the first fusion output module;

[0028] The first fusion output module is used to sum and decode the received input variables to obtain abnormal classification result information.

[0029] The first encoding sub-model includes: a first PReLU activation function module, a first normalization module, a first fully connected layer, a second PReLU activation function module, a second normalization module, a first splicing and fusion module, a second fully connected layer, a third PReLU activation function module, a third fully connected layer, a fourth PReLU activation function module, a third normalization module, a first LSTM network module, a fourth normalization module, a second splicing and fusion module, a fourth fully connected layer, a first Sigmoid activation function module, and a first Hada code inner product module;

[0030] The first PReLU activation function module is used to receive and obtain tooth point cloud data;

[0031] The output end of the first PReLU activation function module is connected to the input end of the first normalization module; the output end of the first normalization module is connected to the input end of the first fully connected layer; the output end of the first fully connected layer is connected to the input end of the second PReLU activation function module; the output end of the second PReLU activation function module is connected to the input end of the second normalization module; the output end of the second normalization module is connected to the input end of the first splicing and fusion module; the output end of the first splicing and fusion module is connected to the input end of the second fully connected layer; the output end of the second fully connected layer is connected to the input end of the third PReLU activation function module; the output end of the third PReLU activation function module is connected to the input end of the third fully connected layer; the output end of the third fully connected layer is connected to the The input end of the fourth PReLU activation function module is connected; the output end of the fourth PReLU activation function module is connected to the input end of the third normalization module; the output end of the third normalization module is connected to the input end of the first LSTM network module; the output end of the first LSTM network module is connected to the input end of the fourth normalization module; the output end of the fourth normalization module is connected to the input end of the second splicing and fusion module; the output end of the second splicing and fusion module is connected to the input end of the fourth fully connected layer; the output end of the fourth fully connected layer is connected to the input end of the first Sigmoid activation function module; the output end of the first Sigmoid activation function module is connected to the input end of the first Hadamard inner product module; the first Hadamard inner product module is used to output a misaligned coded signal.

[0032] The second encoding sub-model includes: a fifth PReLU activation function module, a fifth normalization module, a fifth fully connected layer, a sixth PReLU activation function module, a sixth normalization module, a third splicing and fusion module, a sixth fully connected layer, a seventh PReLU activation function module, a seventh fully connected layer, an eighth PReLU activation function module, a seventh normalization module, a second LSTM network module, an eighth normalization module, a fourth splicing and fusion module, an eighth fully connected layer, a second Sigmoid activation function module, and a second Hadamard inner product module;

[0033] The fifth PReLU activation function module is used to receive and obtain tooth point cloud data;

[0034] The output end of the fifth PReLU activation function module is connected to the input end of the fifth normalization module; the output end of the fifth normalization module is connected to the input end of the fifth fully connected layer; the output end of the fifth fully connected layer is connected to the input end of the sixth PReLU activation function module; the output end of the sixth PReLU activation function module is connected to the input end of the sixth normalization module; the output end of the sixth normalization module is connected to the input end of the third splicing and fusion module; the output end of the third splicing and fusion module is connected to the input end of the sixth fully connected layer; the output end of the sixth fully connected layer is connected to the input end of the seventh PReLU activation function module; the output end of the seventh PReLU activation function module is connected to the input end of the seventh fully connected layer; the output end of the seventh fully connected layer is connected to the The input end of the eighth PReLU activation function module is connected; the output end of the eighth PReLU activation function module is connected to the input end of the seventh normalization module; the output end of the seventh normalization module is connected to the input end of the second LSTM network module; the output end of the second LSTM network module is connected to the input end of the eighth normalization module; the output end of the eighth normalization module is connected to the input end of the fourth splicing and fusion module; the output end of the fourth splicing and fusion module is connected to the input end of the eighth fully connected layer; the output end of the eighth fully connected layer is connected to the input end of the second Sigmoid activation function module; the output end of the second Sigmoid activation function module is connected to the input end of the second Hadamard inner product module; the second Hadamard inner product module is used to output a deformed coding signal.

[0035] The third encoding sub-model includes a first convolution module, a second convolution module, a third convolution module, a fourth convolution module, a first fusion module, a second fusion module, a third fusion module, a first normalization module, a second normalization module, and a first activation module;

[0036] The input end of the first convolution module is used to receive tooth point cloud data, and the output end of the first convolution module is respectively connected to the input end of the first fusion module and the input end of the second fusion module; the output end of the first fusion module is connected to the input end of the third convolution module; the output end of the third convolution module is connected to the input end of the first normalization module; the output end of the first normalization module is connected to the input end of the third fusion module; the input end of the second convolution module is used to receive tooth point cloud data, and the output end of the second convolution module is respectively connected to the input end of the third fusion module and the input end of the first fusion module; the output end of the second fusion module is connected to the input end of the fourth convolution module; the output end of the fourth convolution module is connected to the input end of the second normalization module; the output end of the second normalization module is connected to the input end of the third fusion module; the output end of the third fusion module is connected to the input end of the first activation module; the output end of the first activation module is used to output a missing tooth coding signal.

[0037] The training process of the anomaly recognition model includes:

[0038] Obtaining a dental abnormality classification training data set; the dental abnormality classification training data set includes a plurality of dental abnormality classification training data; the dental abnormality classification training data includes dental point cloud data and corresponding abnormality classification result values;

[0039] Initialize the number of training iterations;

[0040] The tooth point cloud data in the tooth abnormality classification training data set is used as input data and input into the abnormality recognition model; the abnormality classification result value corresponding to the input data is used as label information;

[0041] Using the anomaly recognition model, the input data is processed to obtain anomaly classification result information;

[0042] Perform difference calculation on the obtained abnormal classification result information and label information to obtain a difference value;

[0043] Determine whether the difference value satisfies a convergence condition, and obtain a first determination result;

[0044] When the first judgment result is no, judging whether the training iteration number value is equal to the training number threshold, and obtaining a second judgment result;

[0045] When the second judgment result is no, determining that the model training state does not meet the training termination condition;

[0046] When the second judgment result is yes, determining that the model training state satisfies the training termination condition;

[0047] When the first judgment result is yes, determining that the model training state satisfies a training termination condition;

[0048] When the model training status does not meet the termination training condition, the parameter update model is used to update the parameters of the abnormality recognition model, and the training iteration number value is increased by 1, triggering the execution of inputting the tooth point cloud data in the tooth abnormality classification training data set as input data into the abnormality recognition model;

[0049] When the model training state satisfies the training termination condition, the training process of the anomaly recognition model is completed to obtain a trained anomaly recognition model.

[0050] In a second aspect of the embodiment of the present application, a device for classifying abnormalities in tooth data is disclosed, the device comprising:

[0051] A memory storing executable program code;

[0052] a processor coupled to the memory;

[0053] The processor calls the executable program code stored in the memory to execute the abnormality classification method of the tooth data.

[0054] In a third aspect of the embodiments of the present application, an information data processing terminal is disclosed, and the information data processing terminal is used to implement the abnormality classification method of tooth data.

[0055] The beneficial effects of the present invention are:

[0056] Aiming at the problem of abnormality recognition of dental data, the present invention constructs a special recognition neural network, uses the first coding sub-model, the second coding sub-model, and the third coding sub-model to respectively extract effective abnormal features, and finally uses the fusion model to fuse various features to obtain effective abnormal features, thereby achieving efficient and accurate dental abnormality classification.

[0057] The present invention can realize an orthodontic tooth classification method, can handle various abnormal situations including missing teeth, deformity, dislocation, and adhesion in classification tasks, and has good robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a flow chart for implementing the method of the present invention. DETAILED DESCRIPTION

[0059] In order to better understand the content of the present invention, an embodiment is given here.

[0060] Figure 1 It is a flow chart for implementing the method of the present invention.

[0061] In a first aspect of an embodiment of the present application, a method for classifying abnormalities in tooth data is disclosed, comprising:

[0062] S1, obtaining a tooth point cloud data set; the tooth point cloud data set includes tooth point cloud data;

[0063] S2, preprocessing the tooth point cloud dataset to obtain a preprocessed dataset;

[0064] S3, using the trained abnormality recognition model to perform recognition processing on the preprocessed data set to obtain abnormality classification result information; the values ​​of the abnormality classification result information include normal, missing teeth, deformity, and dislocation.

[0065] The preprocessing of the tooth point cloud data set to obtain a preprocessed data set includes:

[0066] Performing denoising processing on the tooth point cloud dataset to obtain a denoised dataset;

[0067] Performing data cleaning processing on the noise reduction data set to obtain a first data set;

[0068] Performing point cloud registration processing on the first data set to obtain a preprocessed data set;

[0069] The anomaly recognition model includes: a feature fusion sub-model, a first encoding sub-model, a second encoding sub-model, and a third encoding sub-model;

[0070] The first coding sub-model is used to perform dislocation coding processing on the input tooth point cloud data to obtain a dislocation coding signal;

[0071] The second coding sub-model is used to perform deformity coding processing on the input tooth point cloud data to obtain a deformity coding signal;

[0072] The third coding sub-model is used to perform missing tooth coding processing on the input tooth point cloud data to obtain a missing tooth coding signal;

[0073] The feature fusion sub-model is connected to the first coding sub-model, the second coding sub-model, and the third coding sub-model respectively, and is used to fuse the input misaligned coding signals, deformed coding signals, and missing teeth coding signals to obtain abnormal classification result information;

[0074] The feature fusion sub-model includes a first input module, a first self-attention module, a first multi-head attention module, a first addition module, a first fusion output module, a second input module, a second self-attention module, a second multi-head attention module, and a second addition module;

[0075] The first fusion output module is used to output abnormal classification result information;

[0076] The first input module includes an input port, a first output port, and a second output port, and is used to receive the encoding information output by the first encoding sub-model, and output the encoding information through the first output port and the second output port respectively; the first output port of the first input module is connected to the input port of the first self-attention module; the second output port of the first input module is connected to the second input port of the first addition module;

[0077] The first self-attention module includes an input port, a first output port, a second output port, and a third output port; the first output port is used to output the Q variable after the encoding information output by the first encoding sub-model is self-attention encoded; the second output port is used to output the K variable after the encoding information output by the first encoding sub-model is self-attention encoded; the third output port is used to output the V variable after the encoding information output by the first encoding sub-model is self-attention encoded; the first output port of the first self-attention module is connected to the first input port of the first multi-head attention module; the second output port of the first self-attention module is connected to the second input port of the second multi-head attention module; the third output port of the first self-attention module is connected to the third input port of the second multi-head attention module;

[0078] The first multi-head attention module comprises a first input port, a second input port, a third input port and an output port; the output port of the first multi-head attention module is connected to the first input port of the first addition module;

[0079] The first addition module is used to sum the input variable of the first input port and the input variable of the second input port to obtain a sum variable, and send the sum variable to the first fusion output module;

[0080] The second input module includes an input port, a first output port, and a second output port, and is used to receive the encoding information output by the second encoding sub-model, and output the encoding information through the first output port and the second output port respectively; the first output port of the second input module is connected to the input port of the second self-attention module; the second output port of the second input module is connected to the second input port of the first addition module;

[0081] The second self-attention module includes an input port, a first output port, a second output port, and a third output port; the first output port is used to output the Q variable after the encoding information output by the second encoding sub-model is self-attention encoded; the second output port is used to output the K variable after the encoding information output by the second encoding sub-model is self-attention encoded; the third output port is used to output the V variable after the encoding information output by the second encoding sub-model is self-attention encoded; the first output port of the second self-attention module is connected to the first input port of the second multi-head attention module; the second output port of the second self-attention module is connected to the second input port of the second multi-head attention module; the third output port of the second self-attention module is connected to the third input port of the second multi-head attention module;

[0082] The second multi-head attention module comprises a first input port, a second input port, a third input port and an output port; the output port of the second multi-head attention module is connected to the first input port of the second addition module;

[0083] The second addition module is used to sum the input variable of the first input port and the input variable of the second input port to obtain a sum variable, and send the sum variable to the first fusion output module;

[0084] The first fusion output module is used to sum and decode the received input variables to obtain abnormal classification result information;

[0085] The first encoding sub-model includes: a first PReLU activation function module, a first normalization module, a first fully connected layer, a second PReLU activation function module, a second normalization module, a first splicing and fusion module, a second fully connected layer, a third PReLU activation function module, a third fully connected layer, a fourth PReLU activation function module, a third normalization module, a first LSTM network module, a fourth normalization module, a second splicing and fusion module, a fourth fully connected layer, a first Sigmoid activation function module, and a first Hada code inner product module;

[0086] The first PReLU activation function module is used to receive and obtain tooth point cloud data;

[0087] The output end of the first PReLU activation function module is connected to the input end of the first normalization module; the output end of the first normalization module is connected to the input end of the first fully connected layer; the output end of the first fully connected layer is connected to the input end of the second PReLU activation function module; the output end of the second PReLU activation function module is connected to the input end of the second normalization module; the output end of the second normalization module is connected to the input end of the first splicing and fusion module; the output end of the first splicing and fusion module is connected to the input end of the second fully connected layer; the output end of the second fully connected layer is connected to the input end of the third PReLU activation function module; the output end of the third PReLU activation function module is connected to the input end of the third fully connected layer; the output end of the third fully connected layer is connected to the The input end of the fourth PReLU activation function module is connected; the output end of the fourth PReLU activation function module is connected to the input end of the third normalization module; the output end of the third normalization module is connected to the input end of the first LSTM network module; the output end of the first LSTM network module is connected to the input end of the fourth normalization module; the output end of the fourth normalization module is connected to the input end of the second splicing and fusion module; the output end of the second splicing and fusion module is connected to the input end of the fourth fully connected layer; the output end of the fourth fully connected layer is connected to the input end of the first Sigmoid activation function module; the output end of the first Sigmoid activation function module is connected to the input end of the first Hadamard inner product module; the first Hadamard inner product module is used to output a misaligned coded signal.

[0088] The second encoding sub-model includes: a fifth PReLU activation function module, a fifth normalization module, a fifth fully connected layer, a sixth PReLU activation function module, a sixth normalization module, a third splicing and fusion module, a sixth fully connected layer, a seventh PReLU activation function module, a seventh fully connected layer, an eighth PReLU activation function module, a seventh normalization module, a second LSTM network module, an eighth normalization module, a fourth splicing and fusion module, an eighth fully connected layer, a second Sigmoid activation function module, and a second Hadamard inner product module;

[0089] The fifth PReLU activation function module is used to receive and obtain tooth point cloud data;

[0090] The output end of the fifth PReLU activation function module is connected to the input end of the fifth normalization module; the output end of the fifth normalization module is connected to the input end of the fifth fully connected layer; the output end of the fifth fully connected layer is connected to the input end of the sixth PReLU activation function module; the output end of the sixth PReLU activation function module is connected to the input end of the sixth normalization module; the output end of the sixth normalization module is connected to the input end of the third splicing and fusion module; the output end of the third splicing and fusion module is connected to the input end of the sixth fully connected layer; the output end of the sixth fully connected layer is connected to the input end of the seventh PReLU activation function module; the output end of the seventh PReLU activation function module is connected to the input end of the seventh fully connected layer; the output end of the seventh fully connected layer is connected to the The input end of the eighth PReLU activation function module is connected; the output end of the eighth PReLU activation function module is connected to the input end of the seventh normalization module; the output end of the seventh normalization module is connected to the input end of the second LSTM network module; the output end of the second LSTM network module is connected to the input end of the eighth normalization module; the output end of the eighth normalization module is connected to the input end of the fourth splicing and fusion module; the output end of the fourth splicing and fusion module is connected to the input end of the eighth fully connected layer; the output end of the eighth fully connected layer is connected to the input end of the second Sigmoid activation function module; the output end of the second Sigmoid activation function module is connected to the input end of the second Hadamard inner product module; the second Hadamard inner product module is used to output a deformed coding signal.

[0091] The third encoding sub-model includes a first convolution module, a second convolution module, a third convolution module, a fourth convolution module, a first fusion module, a second fusion module, a third fusion module, a first normalization module, a second normalization module, and a first activation module;

[0092] The input end of the first convolution module is used to receive tooth point cloud data, and the output end of the first convolution module is respectively connected to the input end of the first fusion module and the input end of the second fusion module; the output end of the first fusion module is connected to the input end of the third convolution module; the output end of the third convolution module is connected to the input end of the first normalization module; the output end of the first normalization module is connected to the input end of the third fusion module; the input end of the second convolution module is used to receive tooth point cloud data, and the output end of the second convolution module is respectively connected to the input end of the third fusion module and the input end of the first fusion module; the output end of the second fusion module is connected to the input end of the fourth convolution module; the output end of the fourth convolution module is connected to the input end of the second normalization module; the output end of the second normalization module is connected to the input end of the third fusion module; the output end of the third fusion module is connected to the input end of the first activation module; the output end of the first activation module is used to output a missing tooth coding signal.

[0093] The concatenation and fusion module and the fusion module are constructed based on the concat operation in Python; the activation module is implemented based on the sigmoid, or Tanh, or ReLU activation function.

[0094] The first self-attention module is implemented by using the self-attention model in the Transfromer model;

[0095] The first multi-head attention module is implemented using the multi-head attention model in the Transfromer model.

[0096] The training process of the anomaly recognition model includes:

[0097] Obtaining a dental abnormality classification training data set; the dental abnormality classification training data set includes a plurality of dental abnormality classification training data; the dental abnormality classification training data includes dental point cloud data and corresponding abnormality classification result values;

[0098] Initialize the number of training iterations;

[0099] The tooth point cloud data in the tooth abnormality classification training data set is used as input data and input into the abnormality recognition model; the abnormality classification result value corresponding to the input data is used as label information;

[0100] Using the anomaly recognition model, the input data is processed to obtain anomaly classification result information;

[0101] Perform difference calculation on the obtained abnormal classification result information and label information to obtain a difference value;

[0102] Determine whether the difference value satisfies a convergence condition, and obtain a first determination result;

[0103] When the first judgment result is no, judging whether the training iteration number value is equal to the training number threshold, and obtaining a second judgment result;

[0104] When the second judgment result is no, determining that the model training state does not meet the training termination condition;

[0105] When the second judgment result is yes, determining that the model training state satisfies the training termination condition;

[0106] When the first judgment result is yes, determining that the model training state satisfies a training termination condition;

[0107] When the model training status does not meet the termination training condition, the parameter update model is used to update the parameters of the abnormality recognition model, and the training iteration number value is increased by 1, triggering the execution of inputting the tooth point cloud data in the tooth abnormality classification training data set as input data into the abnormality recognition model;

[0108] When the model training state satisfies the training termination condition, the training process of the anomaly recognition model is completed to obtain a trained anomaly recognition model.

[0109] The difference value satisfies the convergence condition, which means that the difference value is less than a preset convergence threshold; the difference value does not satisfy the convergence condition, which means that the difference value is not less than the preset convergence threshold.

[0110] The difference calculation process is implemented by calculating a loss function, and the loss function may adopt a Euclidean distance function or a cross entropy loss function.

[0111] The parameter updating model is expressed as follows:

[0112]

[0113]

[0114] In the formula, is the difference value obtained after the ith tooth point cloud data in the tooth abnormality classification training data set is input into the abnormality recognition model, v is the parameter update value, θ is the parameter of the abnormality recognition model, η is the initial parameter learning rate, α is the momentum angle parameter, 0≤α≤π / 4, It means to find partial derivatives with respect to variable θ;

[0115] The decoding process of the first fusion output module includes: decoding the solution obtained by summing, and when the solution is 0, determining that the value of the abnormal classification result information is normal; when the solution is 1, determining that the value of the abnormal classification result information is missing teeth; when the solution is 2, determining that the value of the abnormal classification result information is deformed; when the solution is 3, determining that the value of the abnormal classification result information is misaligned.

[0116] The noise reduction process can be implemented by using the DeNoise command in MATLAB.

[0117] The point cloud registration process can be implemented by using the FastMAC algorithm or the RegFormer algorithm.

[0118] The data cleaning process includes filling missing values, smoothing noise data, and smoothing or deleting outliers; the smoothed noise data is obtained by first determining the noise data, and then smoothing the noise data according to the data before and after the noise data; the noise data is a value whose value is less than the detection sensitivity of the sensor of the observed data, or greater than the measurement upper limit of the sensor of the observed data. The outlier point can be determined by the Kalman filter method. The filling value of the missing value can be determined by averaging the measured values ​​within a certain sampling interval before and after the missing value.

[0119] The PReLU activation function module is used to implement the PReLU activation function.

[0120] In a second aspect of the embodiment of the present application, a device for classifying abnormalities in tooth data is disclosed, the device comprising:

[0121] A memory storing executable program code;

[0122] a processor coupled to the memory;

[0123] The processor calls the executable program code stored in the memory to execute the abnormality classification method of the tooth data.

[0124] In a third aspect of the embodiments of the present application, an information data processing terminal is disclosed, and the information data processing terminal is used to implement the abnormality classification method of tooth data.

[0125] Aiming at the problem of abnormality recognition of dental data, the present invention constructs a special recognition neural network, uses the first coding sub-model, the second coding sub-model, and the third coding sub-model to respectively extract effective abnormal features, and finally uses the fusion model to fuse various features to obtain effective abnormal features, thereby achieving efficient and accurate dental abnormality classification.

[0126] The present invention can realize an orthodontic tooth classification method, can handle various abnormal situations including missing teeth, deformity, dislocation, and adhesion in classification tasks, and has good robustness.

[0127] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for classifying abnormalities in dental data, characterized in that: include: S1, obtain tooth point cloud dataset; The tooth point cloud data set includes tooth point cloud data; S2, preprocessing the tooth point cloud dataset to obtain a preprocessed dataset; S3, using the trained anomaly recognition model to perform recognition processing on the preprocessed data set to obtain anomaly classification result information; The values ​​of the abnormal classification result information include normal, missing teeth, deformity, and dislocation; The anomaly recognition model includes: a feature fusion sub-model, a first encoding sub-model, a second encoding sub-model, and a third encoding sub-model; The first coding sub-model is used to perform dislocation coding processing on the input tooth point cloud data to obtain a dislocation coding signal; The second coding sub-model is used to perform deformity coding processing on the input tooth point cloud data to obtain a deformity coding signal; The third coding sub-model is used to perform missing tooth coding processing on the input tooth point cloud data to obtain a missing tooth coding signal; The feature fusion sub-model is connected to the first coding sub-model, the second coding sub-model, and the third coding sub-model respectively, and is used to fuse the input misaligned coding signals, deformed coding signals, and missing teeth coding signals to obtain abnormal classification result information; The feature fusion sub-model includes a first input module, a first self-attention module, a first multi-head attention module, a first addition module, a first fusion output module, a second input module, a second self-attention module, a second multi-head attention module, and a second addition module; The first fusion output module is used to output abnormal classification result information; The first input module includes an input port, a first output port, and a second output port, and is used to receive the encoding information output by the first encoding sub-model, and output the encoding information through the first output port and the second output port respectively; the first output port of the first input module is connected to the input port of the first self-attention module; the second output port of the first input module is connected to the second input port of the first addition module; The first self-attention module includes an input port, a first output port, a second output port, and a third output port; the first output port is used to output the Q variable after the encoding information output by the first encoding sub-model is self-attention encoded; the second output port is used to output the K variable after the encoding information output by the first encoding sub-model is self-attention encoded; the third output port is used to output the V variable after the encoding information output by the first encoding sub-model is self-attention encoded; the first output port of the first self-attention module is connected to the first input port of the first multi-head attention module; the second output port of the first self-attention module is connected to the second input port of the second multi-head attention module; the third output port of the first self-attention module is connected to the third input port of the second multi-head attention module; The first multi-head attention module comprises a first input port, a second input port, a third input port and an output port; the output port of the first multi-head attention module is connected to the first input port of the first addition module; The first addition module is used to sum the input variable of the first input port and the input variable of the second input port to obtain a sum variable, and send the sum variable to the first fusion output module; The second input module includes an input port, a first output port, and a second output port, and is used to receive the encoding information output by the second encoding sub-model, and output the encoding information through the first output port and the second output port respectively; the first output port of the second input module is connected to the input port of the second self-attention module; the second output port of the second input module is connected to the second input port of the second addition module; The second self-attention module includes an input port, a first output port, a second output port, and a third output port; the first output port is used to output the Q variable after the encoding information output by the third encoding sub-model is self-attention encoded; the second output port is used to output the K variable after the encoding information output by the third encoding sub-model is self-attention encoded; the third output port is used to output the V variable after the encoding information output by the third encoding sub-model is self-attention encoded; the first output port of the second self-attention module is connected to the first input port of the second multi-head attention module; the second output port of the second self-attention module is connected to the second input port of the second multi-head attention module; the third output port of the second self-attention module is connected to the third input port of the second multi-head attention module; The second multi-head attention module comprises a first input port, a second input port, a third input port and an output port; the output port of the second multi-head attention module is connected to the first input port of the second addition module; The second addition module is used to sum the input variable of the first input port and the input variable of the second input port to obtain a sum variable, and send the sum variable to the first fusion output module; The first fusion output module is used to sum and decode the received input variables to obtain abnormal classification result information.

2. The method for classifying abnormalities in dental data according to claim 1, characterized in that: The preprocessing of the tooth point cloud data set to obtain a preprocessed data set includes: S21, performing noise reduction processing on the tooth point cloud dataset to obtain a noise reduction dataset; S22, performing data cleaning processing on the noise reduction data set to obtain a first data set; S23, performing point cloud registration processing on the first data set to obtain a preprocessed data set.

3. The method for abnormal classification of dental data according to claim 1, characterized in that: The first encoding sub-model includes: a first PReLU activation function module, a first normalization module, a first fully connected layer, a second PReLU activation function module, a second normalization module, a first splicing and fusion module, a second fully connected layer, a third PReLU activation function module, a third fully connected layer, a fourth PReLU activation function module, a third normalization module, a first LSTM network module, a fourth normalization module, a second splicing and fusion module, a fourth fully connected layer, a first Sigmoid activation function module, and a first Hada code inner product module; The first PReLU activation function module is used to receive and obtain tooth point cloud data; The output end of the first PReLU activation function module is connected to the input end of the first normalization module; the output end of the first normalization module is connected to the input end of the first fully connected layer; the output end of the first fully connected layer is connected to the input end of the second PReLU activation function module; the output end of the second PReLU activation function module is connected to the input end of the second normalization module; the output end of the second normalization module is connected to the input end of the first splicing and fusion module; the output end of the first splicing and fusion module is connected to the input end of the second fully connected layer; the output end of the second fully connected layer is connected to the input end of the third PReLU activation function module; the output end of the third PReLU activation function module is connected to the input end of the third fully connected layer; the output end of the third fully connected layer is connected to the The input end of the fourth PReLU activation function module is connected; the output end of the fourth PReLU activation function module is connected to the input end of the third normalization module; the output end of the third normalization module is connected to the input end of the first LSTM network module; the output end of the first LSTM network module is connected to the input end of the fourth normalization module; the output end of the fourth normalization module is connected to the input end of the second splicing and fusion module; the output end of the second splicing and fusion module is connected to the input end of the fourth fully connected layer; the output end of the fourth fully connected layer is connected to the input end of the first Sigmoid activation function module; the output end of the first Sigmoid activation function module is connected to the input end of the first Hadamard inner product module; the first Hadamard inner product module is used to output a misaligned coded signal.

4. The method for abnormal classification of dental data according to claim 1, characterized in that: The second encoding sub-model includes: a fifth PReLU activation function module, a fifth normalization module, a fifth fully connected layer, a sixth PReLU activation function module, a sixth normalization module, a third splicing and fusion module, a sixth fully connected layer, a seventh PReLU activation function module, a seventh fully connected layer, an eighth PReLU activation function module, a seventh normalization module, a second LSTM network module, an eighth normalization module, a fourth splicing and fusion module, an eighth fully connected layer, a second Sigmoid activation function module, and a second Hadamard inner product module; The fifth PReLU activation function module is used to receive and obtain tooth point cloud data; The output end of the fifth PReLU activation function module is connected to the input end of the fifth normalization module; the output end of the fifth normalization module is connected to the input end of the fifth fully connected layer; the output end of the fifth fully connected layer is connected to the input end of the sixth PReLU activation function module; the output end of the sixth PReLU activation function module is connected to the input end of the sixth normalization module; the output end of the sixth normalization module is connected to the input end of the third splicing and fusion module; the output end of the third splicing and fusion module is connected to the input end of the sixth fully connected layer; the output end of the sixth fully connected layer is connected to the input end of the seventh PReLU activation function module; the output end of the seventh PReLU activation function module is connected to the input end of the seventh fully connected layer; the output end of the seventh fully connected layer is connected to the The input end of the eighth PReLU activation function module is connected; the output end of the eighth PReLU activation function module is connected to the input end of the seventh normalization module; the output end of the seventh normalization module is connected to the input end of the second LSTM network module; the output end of the second LSTM network module is connected to the input end of the eighth normalization module; the output end of the eighth normalization module is connected to the input end of the fourth splicing and fusion module; the output end of the fourth splicing and fusion module is connected to the input end of the eighth fully connected layer; the output end of the eighth fully connected layer is connected to the input end of the second Sigmoid activation function module; the output end of the second Sigmoid activation function module is connected to the input end of the second Hadamard inner product module; the second Hadamard inner product module is used to output a deformed coding signal.

5. The method for abnormal classification of dental data according to claim 1, characterized in that: The third encoding sub-model includes a first convolution module, a second convolution module, a third convolution module, a fourth convolution module, a first fusion module, a second fusion module, a third fusion module, a first normalization module, a second normalization module, and a first activation module; The input end of the first convolution module is used to receive tooth point cloud data, and the output end of the first convolution module is respectively connected to the input end of the first fusion module and the input end of the second fusion module; the output end of the first fusion module is connected to the input end of the third convolution module; the output end of the third convolution module is connected to the input end of the first normalization module; the output end of the first normalization module is connected to the input end of the third fusion module; the input end of the second convolution module is used to receive tooth point cloud data, and the output end of the second convolution module is respectively connected to the input end of the third fusion module and the input end of the first fusion module; the output end of the second fusion module is connected to the input end of the fourth convolution module; the output end of the fourth convolution module is connected to the input end of the second normalization module; the output end of the second normalization module is connected to the input end of the third fusion module; the output end of the third fusion module is connected to the input end of the first activation module; the output end of the first activation module is used to output a missing tooth coding signal.

6. The method for abnormal classification of dental data according to claim 1, characterized in that: The training process of the anomaly recognition model includes: Obtaining a dental abnormality classification training data set; the dental abnormality classification training data set includes a plurality of dental abnormality classification training data; the dental abnormality classification training data includes dental point cloud data and corresponding abnormality classification result values; Initialize the number of training iterations; The tooth point cloud data in the tooth abnormality classification training data set is used as input data and input into the abnormality recognition model; the abnormality classification result value corresponding to the input data is used as label information; Using the anomaly recognition model, the input data is processed to obtain anomaly classification result information; Perform difference calculation on the obtained abnormal classification result information and label information to obtain a difference value; Determine whether the difference value satisfies a convergence condition, and obtain a first determination result; When the first judgment result is no, judging whether the training iteration number value is equal to the training number threshold, and obtaining a second judgment result; When the second judgment result is no, determining that the model training state does not meet the training termination condition; When the second judgment result is yes, determining that the model training state satisfies the training termination condition; When the first judgment result is yes, determining that the model training state satisfies a training termination condition; When the model training status does not meet the termination training condition, the parameter update model is used to update the parameters of the abnormality recognition model, and the training iteration number value is increased by 1, triggering the execution of inputting the tooth point cloud data in the tooth abnormality classification training data set as input data into the abnormality recognition model; When the model training state satisfies the training termination condition, the training process of the anomaly recognition model is completed to obtain a trained anomaly recognition model.

7. A device for classifying abnormalities in dental data, characterized in that: The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the abnormality classification method for dental data according to any one of claims 1 to 6.

8. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the abnormality classification method of tooth data according to any one of claims 1 to 6.

Images