Identification processing method and device for oral dentition video data

Abstract

The invention discloses an identification processing method and device for oral dentition video data. The method comprises the following steps: S1, acquiring a depth video data set of an oral dentition; the depth video data set comprises color video data and depth video data; s2, performing point cloud data construction processing on the depth video data set to obtain a point cloud data set; and S3, performing identification processing on the point cloud data set to obtain user identification information.

Description

Technical Field

[0001] This invention relates to the fields of intelligent video data processing and point cloud data modeling processing, specifically to a method and apparatus for recognizing and processing oral dental arch video data. Background Technology

[0002] In the field of oral medicine, accurate identification and analysis of dentition are crucial for applications such as orthodontics and restorations. Traditional methods for dentition detection primarily rely on two-dimensional images or manual measurements, which present the following technical challenges:

[0003] 1. Limitations of 2D images: 2D images cannot provide depth information about teeth, leading to inaccurate analysis of tooth shape and position. For example, in cases of complex tooth arrangement or overlap, 2D images struggle to distinguish the three-dimensional structure of teeth.

[0004] 2. Errors in manual measurement: Manual measurement methods rely on the doctor's experience and skills, and are easily affected by subjective factors, resulting in poor repeatability and accuracy of measurement results.

[0005] 3. Lack of automation and efficiency: Traditional methods usually require a lot of manual operation, which is time-consuming and labor-intensive, and it is difficult to meet the needs of large-scale clinical applications.

[0006] 4. Challenges of multi-view data fusion: In practical applications, it is usually necessary to collect dental data from multiple perspectives, but existing technologies have difficulties in multi-view data fusion, making it difficult to effectively integrate depth and color information from different perspectives.

[0007] 5. Insufficient tooth segmentation accuracy: Existing point cloud segmentation algorithms are prone to missegmentation or omission when dealing with complex dental structures, leading to a decrease in the accuracy of subsequent recognition and analysis. Summary of the Invention

[0008] This invention mainly addresses the problems of low efficiency and poor accuracy in existing tooth recognition methods. This invention discloses a method and apparatus for recognizing and processing oral dental video data.

[0009] In a first aspect, the present invention discloses a method for recognizing and processing oral dental arch video data, characterized in that it includes:

[0010] S1, Acquire a set of depth video data of the dental arch; the set of depth video data includes color video data and depth video data;

[0011] S2, perform point cloud data construction processing on the depth video data set to obtain the total point cloud dataset;

[0012] S3, perform recognition processing on the total point cloud dataset to obtain user identification information.

[0013] The point cloud data construction process performed on the depth video data set to obtain the total point cloud dataset includes:

[0014] S21, extract a depth image set from the depth video data in the depth video data set; the depth image set includes depth images;

[0015] S22, extract a color image set from the color video data in the color video data set; the color image set includes color images;

[0016] S23, perform image matching processing on the depth image set and the color image set to obtain a matching image set; the matching image set includes several matching depth images and color images;

[0017] S24. Based on the matching image set, the total point cloud dataset is constructed.

[0018] The process of identifying the total point cloud dataset to obtain user identification information includes:

[0019] S31, preprocess the total point cloud dataset to obtain a preprocessed dataset;

[0020] S32, perform point cloud segmentation processing on the preprocessed dataset to obtain the point cloud dataset for each tooth;

[0021] S33, perform recognition processing on the point cloud dataset of all teeth to obtain user identification information; the user identification information is used to characterize the user information to which the depth video data set of the oral dentition belongs.

[0022] The preprocessing of the total point cloud dataset to obtain a preprocessed dataset includes:

[0023] The total point cloud dataset is denoised to obtain the first dataset;

[0024] The first dataset is filtered to obtain a preprocessed dataset.

[0025] The step of performing point cloud segmentation on the preprocessed dataset to obtain a point cloud dataset for each tooth includes:

[0026] S321, using the first segmentation algorithm, perform point cloud segmentation processing on the preprocessed dataset to obtain the first segmentation result information set;

[0027] S322, using the second segmentation algorithm, point cloud segmentation processing is performed on the preprocessed dataset to obtain the second segmentation result information set;

[0028] S323, using the third segmentation algorithm, point cloud segmentation processing is performed on the preprocessed dataset to obtain the third segmentation result information set; the first, second and third segmentation result information sets all include the point cloud dataset of each tooth;

[0029] S324, perform fusion and classification processing on all segmentation result information sets to obtain the final segmentation result information set; the final segmentation result information set includes the point cloud dataset of each tooth obtained after fusion and classification processing.

[0030] The process of fusing and classifying all segmentation result information sets to obtain the final segmentation result information set includes:

[0031] S3241, Obtain the point cloud dataset of the same teeth from the first, second and third segmentation result information sets;

[0032] S3242, Statistical parameters are calculated for the point cloud datasets of the same teeth in the first, second and third segmentation result information sets respectively to obtain the corresponding statistical parameter values; The statistical parameter values ​​include the mean, variance, transform domain mean and transform domain variance of all data in a point cloud dataset;

[0033] S3243, using all statistical parameter values, perform fusion calculation on the point cloud datasets of the same teeth in the first, second and third segmentation result information sets to obtain the corresponding point cloud dataset of teeth after fusion classification processing;

[0034] S3244: Using the point cloud dataset of teeth obtained after all the fusion and classification processes, construct the final segmentation result information set.

[0035] The expression for the fusion calculation is:

[0036]

[0037] Where, d 1i d 2i d 3i Let be the i-th data point of the point cloud dataset of the same tooth in the first, second, and third segmentation result information sets, respectively, and αj1, αj2, βj1, and βj2 be the mean, variance, transform domain mean, and transform domain variance of all data points of the point cloud dataset of the same tooth in the j-th segmentation result information set, respectively, where j = 1, 2, and 3. N1 represents the i-th data point in the point cloud dataset of teeth obtained after the point cloud dataset has been fused and classified. N1 represents the total number of data points in the point cloud dataset.

[0038] A second aspect of the present invention discloses a device for recognizing and processing oral dental arch video data, the device comprising:

[0039] Memory containing executable program code;

[0040] A processor coupled to the memory;

[0041] The processor calls the executable program code stored in the memory to execute the oral dental video data recognition and processing method.

[0042] In a third aspect of this invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions, and when the computer instructions are invoked by a computer, they are used to execute the method for recognizing and processing oral dental video data.

[0043] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the method for recognizing and processing oral dental video data.

[0044] The beneficial effects of this invention are as follows:

[0045] This invention, through the fusion of depth and color video data, can accurately reconstruct a 3D point cloud model of the dental arch, providing depth and color information for the teeth and solving the problem that 2D images cannot provide depth information. Utilizing multi-view data fusion technology, it can effectively integrate depth and color information from different perspectives to generate a more complete and accurate 3D model of the dental arch. Multiple segmentation algorithms (such as K-means, BIRCH, and DBSCAN) are used to segment the point cloud data, and through fusion classification processing, the accuracy and robustness of tooth segmentation are improved, solving the problem of insufficient tooth segmentation accuracy in existing technologies. The fusion calculation method further optimizes the segmentation effect by integrating statistical parameters from multiple segmentation results, avoiding the limitations of a single algorithm.

[0046] This invention automates the entire process from data acquisition to tooth identification, reducing manual intervention and improving efficiency and result stability. Through generalized feature value decomposition and recognition fusion calculation, it can quickly and accurately identify the user's dental arch, making it suitable for large-scale clinical applications. The precise three-dimensional dental arch model and efficient identification technology provided by this invention offer dentists more accurate diagnostic information, helping to develop more scientific treatment plans and improve treatment outcomes. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating the implementation of the method of the present invention. Detailed Implementation

[0048] To better understand the content of this invention, an embodiment is provided here.

[0049] Figure 1 This is a flowchart illustrating the implementation of the method of the present invention.

[0050] In a first aspect, this invention discloses a method for recognizing and processing oral dental arch video data, comprising:

[0051] S1, Acquire a set of depth video data of the dental arch; the set of depth video data includes color video data and depth video data;

[0052] S2, perform point cloud data construction processing on the depth video data set to obtain the total point cloud dataset;

[0053] S3, perform recognition processing on the point cloud dataset to obtain user identification information.

[0054] The point cloud data construction process performed on the depth video data set to obtain the total point cloud dataset includes:

[0055] S21, extract a depth image set from the depth video data in the depth video data set; the depth image set includes depth images;

[0056] S22, extract a color image set from the color video data in the color video data set; the color image set includes color images;

[0057] S23, perform image matching processing on the depth image set and the color image set to obtain a matching image set; the matching image set includes several matching depth images and color images;

[0058] S24. Based on the matching image set, the total point cloud dataset is constructed;

[0059] The image extracted from the depth video data or color video data can be obtained using a video frame extraction method.

[0060] The process of identifying the total point cloud dataset to obtain user identification information includes:

[0061] S31, preprocess the total point cloud dataset to obtain a preprocessed dataset;

[0062] S32, perform point cloud segmentation processing on the preprocessed dataset to obtain the point cloud dataset for each tooth;

[0063] S33, perform recognition processing on the point cloud dataset of all teeth to obtain user identification information; the user identification information is used to characterize the user information to which the depth video data set of the oral dentition belongs;

[0064] The preprocessing of the total point cloud dataset to obtain a preprocessed dataset includes:

[0065] The total point cloud dataset is denoised to obtain the first dataset;

[0066] The first dataset is filtered to obtain a preprocessed dataset;

[0067] Both the noise reduction and filtering processes can employ filtering and noise reduction methods used in point cloud data processing.

[0068] The step of performing point cloud segmentation on the preprocessed dataset to obtain a point cloud dataset for each tooth includes:

[0069] S321, using the first segmentation algorithm, perform point cloud segmentation processing on the preprocessed dataset to obtain the first segmentation result information set;

[0070] S322, using the second segmentation algorithm, point cloud segmentation processing is performed on the preprocessed dataset to obtain the second segmentation result information set;

[0071] S323, using the third segmentation algorithm, point cloud segmentation processing is performed on the preprocessed dataset to obtain the third segmentation result information set; the first, second and third segmentation result information sets all include the point cloud dataset of each tooth;

[0072] S324, perform fusion and classification processing on all segmentation result information sets to obtain the final segmentation result information set; the final segmentation result information set includes the point cloud dataset of each tooth obtained after fusion and classification processing.

[0073] The process of fusing and classifying all segmentation result information sets to obtain the final segmentation result information set includes:

[0074] S3241, Obtain the point cloud dataset of the same teeth from the first, second and third segmentation result information sets;

[0075] S3242, Statistical parameters are calculated for the point cloud datasets of the same teeth in the first, second and third segmentation result information sets respectively to obtain the corresponding statistical parameter values; The statistical parameter values ​​include the mean, variance, transform domain mean and transform domain variance of all data in a point cloud dataset;

[0076] S3243, using all statistical parameter values, perform fusion calculation on the point cloud datasets of the same teeth in the first, second and third segmentation result information sets to obtain the corresponding point cloud dataset of teeth after fusion classification processing;

[0077] S3244: Using the point cloud dataset of teeth obtained after all the fusion and classification processes, construct the final segmentation result information set.

[0078] The expression for the fusion calculation is:

[0079]

[0080] Where, d 1i d 2i d 3i Let be the i-th data point of the point cloud dataset of the same tooth in the first, second, and third segmentation result information sets, respectively, and αj1, αj2, βj1, and βj2 be the mean, variance, transform domain mean, and transform domain variance of all data points of the point cloud dataset of the same tooth in the j-th segmentation result information set, respectively, where j = 1, 2, and 3. N1 represents the i-th data point in the point cloud dataset of teeth obtained after the point cloud dataset has been fused and classified. N1 represents the total number of data points in the point cloud dataset.

[0081] This formula integrates the results of three different segmentation algorithms (such as K-means, BIRCH, and DBSCAN), using the statistical parameters (mean, variance, transform domain mean, and transform domain variance) of each algorithm for weighted fusion, effectively compensating for the shortcomings of a single algorithm. For example, the K-means algorithm performs well in handling regular shapes but is sensitive to noise; the DBSCAN algorithm has good robustness to noise but may perform poorly in certain boundary cases. Through fusion calculation, the advantages of each algorithm can be fully utilized, improving the accuracy and robustness of the segmentation results. The weights in the formula are dynamically calculated based on the statistical parameters of each segmentation result, automatically adjusting the contribution of each algorithm according to the distribution characteristics of the data. This adaptive weight allocation mechanism makes the fusion result more reasonable and avoids segmentation bias caused by changes in data characteristics. The transform domain mean and variance obtained through Feature Pattern Decomposition (REMD) further enhance the ability to characterize the intrinsic structure of point cloud data. This multi-dimensional feature fusion can more accurately identify the boundaries of each tooth, thereby improving the accuracy of tooth segmentation and providing high-quality input data for subsequent recognition processing.

[0082] The transform domain mean and transform domain variance are obtained by performing Feature Pattern Decomposition (REMD) on all data in a point cloud dataset to obtain transform domain data, and then calculating the mean and variance of all transform domain data.

[0083] The point cloud datasets of the same teeth obtained from the first, second, and third segmentation result information sets can be used as the point cloud datasets of the same teeth in the three segmentation result information sets, based on the distance information of each point cloud dataset.

[0084] The first segmentation algorithm, the second segmentation algorithm, and the third segmentation algorithm can be the K-means algorithm, the BIRCH algorithm, and the DBSCAN algorithm, respectively.

[0085] The point cloud dataset of all teeth is processed to obtain user identification information, including:

[0086] Obtain the standard point cloud dataset of teeth for all users to be identified; the standard point cloud dataset of teeth includes the standard point cloud dataset of each tooth;

[0087] The point cloud dataset of all teeth is represented as a tooth point cloud data matrix; the row vector of the tooth point cloud data matrix is ​​the point cloud dataset of one tooth.

[0088] The elements of the row vector are data points from a point cloud dataset of a tooth;

[0089] The tooth point cloud data matrix and the standard tooth point cloud dataset of all users to be identified are processed for recognition calculation to obtain the recognition classification value of each user to be identified.

[0090] Identify the user to be identified corresponding to the minimum recognition classification value, and then identify the user information.

[0091] The process of performing recognition calculations on the tooth point cloud data matrix and the standard tooth point cloud datasets of all users to be identified to obtain the recognition classification value for each user includes:

[0092] The standard point cloud dataset of each user's teeth is represented as a corresponding standard matrix; the row vector of the standard matrix is ​​the point cloud dataset of one tooth.

[0093] Subtract the standard matrix from the tooth point cloud data matrix to obtain the point cloud difference matrix;

[0094] Perform pole-symmetric mode decomposition on each row vector of the point cloud difference matrix to obtain the corresponding transformation vector;

[0095] By using all the transformation vectors as row vectors, we can construct the transformation matrix.

[0096] Using the point cloud difference matrix and transformation matrix as a matrix bundle, the matrix bundle is subjected to generalized eigenvalue decomposition to obtain generalized eigenvalues ​​and generalized eigenvectors.

[0097] The recognition and fusion calculation is performed on all generalized feature values ​​and generalized feature vectors to obtain the recognition classification value of the user to be identified.

[0098] The expression for the recognition fusion calculation is:

[0099]

[0100] Where f is the classification value of the user to be identified, and a ij γ is the j-th element of the i-th generalized eigenvector. i Let a be the i-th generalized eigenvalue. i Let N be the mean of the i-th generalized eigenvector, N be the number of elements in the generalized eigenvector, and M be the number of generalized eigenvalues.

[0101] The expression for the recognition fusion calculation comprehensively analyzes the point cloud difference matrix and transformation matrix using generalized eigenvalues ​​and generalized eigenvectors obtained from generalized eigenvalue decomposition. This multi-feature fusion method fully utilizes the global and local characteristics of point cloud data to improve recognition accuracy. For example, generalized eigenvalues ​​reflect the main trends in point cloud data, while generalized eigenvectors provide detailed information. By combining the two, the shape and positional features of teeth can be more comprehensively characterized. The arctan function and absolute value operation in the formula effectively suppress the influence of outliers, enhancing the robustness of the recognition process. In practical applications, tooth point cloud data may be affected by noise, occlusion, and other factors; this robust design ensures accurate identification of user information even under complex conditions. Through recognition fusion calculation, the tooth point cloud data of the user to be identified can be precisely matched with standard point cloud data to determine the user corresponding to the minimum recognition classification value. This method not only quickly identifies users but also provides high-precision classification results, making it suitable for user identification and dentition analysis in large-scale clinical applications. The formula is designed to take into account the differences in tooth structure among different users. Through generalized feature value decomposition and fusion calculation, it can adapt to various tooth shapes and arrangements. This adaptability and generalization ability enable the recognition method of this invention to maintain high accuracy and reliability in different application scenarios.

[0102] The image matching processing of the depth image set and the color image set can employ SIFT (Scale Invariant Feature Transform), SURF (Speed ​​Robust Feature Transform), and ORB (Oriented Fast and Rotated BRIEF) algorithms.

[0103] The process of constructing a total point cloud dataset based on the matched image set includes: conversion from pixel coordinates to camera coordinates, conversion of the depth value of each pixel in the depth image to point coordinates in three-dimensional space, calculation of point coordinates in the camera coordinate system, conversion from camera coordinate system to world coordinate system, conversion from depth value to point cloud, coloring of point cloud, filtering and optimization of point cloud, stitching and fusion of point cloud, and post-processing of point cloud.

[0104] The filtering process can employ interpolation or morphological methods.

[0105] A second aspect of the present invention discloses a device for recognizing and processing oral dental arch video data, the device comprising:

[0106] Memory containing executable program code;

[0107] A processor coupled to the memory;

[0108] The processor calls the executable program code stored in the memory to execute the oral dental video data recognition and processing method.

[0109] In a third aspect of this invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions, and when the computer instructions are invoked by a computer, they are used to execute the method for recognizing and processing oral dental video data.

[0110] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the method for recognizing and processing oral dental video data.

[0111] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for recognizing and processing oral dental arch video data, characterized in that, include: S1, Acquire a set of depth video data of the dental arch; the set of depth video data includes color video data and depth video data; S2, perform point cloud data construction processing on the depth video data set to obtain the total point cloud dataset; S3, perform recognition processing on the total point cloud dataset to obtain user identification information.

2. The method for recognizing and processing oral dental arch video data as described in claim 1, characterized in that, The point cloud data construction process performed on the depth video data set to obtain the total point cloud dataset includes: S21, extract a depth image set from the depth video data in the depth video data set; the depth image set includes depth images; S22, extract a color image set from the color video data in the color video data set; the color image set includes color images; S23, perform image matching processing on the depth image set and the color image set to obtain a matching image set; the matching image set includes several matching depth images and color images; S24. Based on the matching image set, the total point cloud dataset is constructed.

3. The method for recognizing and processing oral dental arch video data as described in claim 1, characterized in that, The process of identifying the total point cloud dataset to obtain user identification information includes: S31, preprocess the total point cloud dataset to obtain a preprocessed dataset; S32, perform point cloud segmentation processing on the preprocessed dataset to obtain the point cloud dataset for each tooth; S33, perform recognition processing on the point cloud dataset of all teeth to obtain user identification information; the user identification information is used to characterize the user information to which the depth video data set of the oral dentition belongs.

4. The method for recognizing and processing oral dental arch video data as described in claim 3, characterized in that, The preprocessing of the total point cloud dataset to obtain a preprocessed dataset includes: The total point cloud dataset is denoised to obtain the first dataset; The first dataset is filtered to obtain a preprocessed dataset.

5. The method for recognizing and processing oral dental arch video data as described in claim 3, characterized in that, The step of performing point cloud segmentation on the preprocessed dataset to obtain a point cloud dataset for each tooth includes: S321, using the first segmentation algorithm, perform point cloud segmentation processing on the preprocessed dataset to obtain the first segmentation result information set; S322, using the second segmentation algorithm, point cloud segmentation processing is performed on the preprocessed dataset to obtain the second segmentation result information set; S323, using the third segmentation algorithm, perform point cloud segmentation processing on the preprocessed dataset to obtain a third segmentation result information set; the first, second and third segmentation result information sets all include the point cloud dataset of each tooth; S324, perform fusion and classification processing on all segmentation result information sets to obtain the final segmentation result information set; the final segmentation result information set includes the point cloud dataset of each tooth obtained after fusion and classification processing.

6. The method for recognizing and processing oral dental arch video data as described in claim 5, characterized in that, The process of fusing and classifying all segmentation result information sets to obtain the final segmentation result information set includes: S3241, Obtain the point cloud dataset of the same teeth from the first, second and third segmentation result information sets; S3242, Statistical parameters are calculated for the point cloud datasets of the same teeth in the first, second and third segmentation result information sets respectively to obtain the corresponding statistical parameter values; The statistical parameter values ​​include the mean, variance, transform domain mean and transform domain variance of all data in a point cloud dataset; S3243, using all statistical parameter values, perform fusion calculation on the point cloud datasets of the same teeth in the first, second and third segmentation result information sets to obtain the corresponding point cloud dataset of teeth after fusion classification processing; S3244: Using the point cloud dataset of teeth obtained after all the fusion and classification processes, a final segmentation result information set is constructed.

7. The method for recognizing and processing oral dental arch video data as described in claim 6, characterized in that, The expression for the fusion calculation is: Where, d 1i d 2i d 3i Let be the i-th data point of the point cloud dataset of the same tooth in the first, second, and third segmentation result information sets, respectively, and αj1, αj2, βj1, and βj2 be the mean, variance, transform domain mean, and transform domain variance of all data points of the point cloud dataset of the same tooth in the j-th segmentation result information set, respectively, where j = 1, 2, and 3. N1 represents the i-th data point in the point cloud dataset of teeth obtained after the point cloud dataset has been fused and classified. N1 represents the total number of data points in the point cloud dataset.

8. A device for recognizing and processing oral dental arch video data, characterized in that, The device includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the method for recognizing and processing oral dental video data as described in any one of claims 1 to 7.

9. A computer-storable medium, characterized in that, The computer storage medium stores computer instructions, which, when invoked by the computer, are used to execute the method for recognizing and processing oral dental video data as described in any one of claims 1 to 7.

10. An information data processing terminal, characterized in that, The information data processing terminal is used to implement the method for recognizing and processing oral dental video data as described in any one of claims 1 to 7.

Images