A recognition processing method and device based on multi-angle front tooth photos

Abstract

The application discloses a kind of based on multi-angle front tooth photo identification processing method and device, the method includes: the image information set of user front tooth is obtained;The image information set includes multiple angle image subset;The angle image subset is the image information set of user front tooth obtained from an angle photographing;The image information set is preprocessed, and tooth image set is obtained;The tooth image set includes the image subset of each tooth;The tooth image set is morphologically identified and processed, and morphological identification result information set is obtained;The morphological identification result information set includes the morphological identification result information of each tooth.

Description

Technical Field

[0001] This invention relates to the fields of oral image data processing and big data recognition and modeling processing, specifically to a recognition and processing method and apparatus based on multi-angle anterior teeth photographs. Background Technology

[0002] In the field of oral medicine, the morphological identification and analysis of anterior teeth is of great significance. Traditional methods for anterior tooth morphological identification mainly rely on the subjective judgment and experience of dentists, assessing tooth morphological characteristics by observing patient oral models or two-dimensional X-rays. However, these methods have the following technical problems:

[0003] 1. Highly subjective, with poor accuracy and consistency: Doctors' experience and subjective judgments vary from person to person, which may lead to significant differences in the morphological identification results of the same tooth.

[0004] 2. Limited information in two-dimensional images: Two-dimensional images such as X-rays cannot fully reflect the three-dimensional morphological characteristics of teeth and are prone to missing some key information, such as the rotation angle of the teeth and the labial and lingual morphology, thus limiting the accurate analysis of tooth morphology.

[0005] 3. Limited accuracy in tooth segmentation: Teeth are closely packed in the oral cavity and vary in shape and size. Existing segmentation algorithms struggle to accurately separate each tooth from an image, easily leading to missegmentation or omissions. For example, when the gaps between adjacent teeth are small, or when there are defects such as cavities on the tooth surface, the segmentation algorithm may misclassify adjacent teeth as a single unit or miss a portion of the tooth.

[0006] 4. Insufficient generalization ability of morphological recognition: Existing morphological recognition methods usually rely on specific image features or template matching, which has poor adaptability to different patients, different shooting angles, and different tooth morphologies, making it difficult to accurately identify various complex situations. For example, when teeth have non-standard morphologies such as wear or rotation, recognition methods based on fixed templates may fail. Summary of the Invention

[0007] This invention mainly addresses the problems of limited tooth segmentation accuracy and insufficient generalization ability of morphological recognition in existing anterior tooth recognition methods. This invention discloses a recognition and processing method and device based on multi-angle anterior tooth photographs.

[0008] In a first aspect, this invention discloses a method for identifying and processing anterior teeth based on multi-angle anterior teeth photographs, comprising:

[0009] S1, acquire a set of image information of the user's anterior teeth; the set of image information includes multiple subsets of images from different angles; the subset of images from different angles is a set of image information of the user's anterior teeth captured from one angle;

[0010] S2, preprocess the image information set to obtain a set of tooth images; the set of tooth images includes a subset of images for each tooth;

[0011] S3, perform morphological recognition processing on the set of tooth images to obtain a set of morphological recognition result information; the set of morphological recognition result information includes the morphological recognition result information of each tooth.

[0012] The preprocessing of the image information set to obtain a set of tooth images includes:

[0013] S21, perform noise reduction processing on the image information set to obtain a first image set;

[0014] S22, perform image enhancement processing on the first image set to obtain the second image set;

[0015] S23, perform tooth segmentation processing on the second image set to obtain a tooth image set.

[0016] The step of performing image enhancement processing on the first image set to obtain the second image set includes:

[0017] For each image in the first image set, a grayscale histogram is calculated to obtain the corresponding grayscale histogram;

[0018] Perform gray-level value statistical processing on the gray-level histogram to obtain the gray-level probability density function;

[0019] The gray-level probability density function is inverted to obtain the inverse gray-level probability density function;

[0020] The gray-level inverse probability density function is subjected to integral enhancement transformation to obtain the enhancement function;

[0021] Using the enhancement function, the pixels of each image in the first image set are calculated to obtain the corresponding enhanced pixels. Using all the enhanced pixels, the corresponding enhanced image is constructed.

[0022] Using all the enhanced images, a second image set is constructed.

[0023] The step of performing tooth segmentation processing on the second image set to obtain a tooth image set includes:

[0024] S231, using a preset dental segmentation network, perform tooth segmentation processing on the second image set to obtain a first recognition result set; the first recognition result set includes the first sequence number information of teeth in the sub-image of each second image in the second image set;

[0025] S232, perform image template matching and recognition processing on the second image set to obtain a second recognition result set; the second recognition result set includes the second sequence number information of the teeth of each sub-image of the second image in the second image set;

[0026] S233, perform a first fusion calculation on the first recognition result set and the second recognition result set to obtain a set of tooth images.

[0027] The first fusion calculation of the first recognition result set and the second recognition result set to obtain a set of tooth images includes:

[0028] S2331, For each sub-image in the first recognition result set and the second recognition result set, calculate the mean, variance and median value of all pixel gray values ​​of the sub-image respectively;

[0029] S2332, For the same sub-image in the first recognition result set and the second recognition result set, perform pixel value fusion and weighted calculation on the corresponding first tooth number information and second tooth number information to obtain the corresponding tooth number information;

[0030] S2333, using all sub-images corresponding to the same tooth serial number information, construct a subset of images of the teeth corresponding to the tooth serial number information;

[0031] S2334, using all the tooth image subsets, construct a tooth image set.

[0032] The expression for the pixel value fusion weighted calculation is:

[0033]

[0034] Where α1, β1, and ∈1 are the mean, variance, and median of all pixel gray values ​​in the sub-images of the first recognition result set, respectively; α2, β2, and ∈2 are the mean, variance, and median of all pixel gray values ​​in the sub-images of the second recognition result set, respectively; and N1 and N2 are the first and second sequence numbers of the teeth, respectively. and These represent rounding down and rounding up, respectively. N0 represents the tooth number information, and T2() is the second-order Legendre polynomial.

[0035] The morphological recognition processing of the tooth image set to obtain a morphological recognition result information set includes:

[0036] Using the image subset of each tooth in the tooth image set, an image tensor corresponding to the tooth is constructed; the image tensor corresponding to the tooth is a three-dimensional tensor; the elements of the three-dimensional tensor with dimensions i1, j1 and k1 are the gray values ​​of the pixels in the j1th row and k1th column of the i1th sub-image of the tooth image subset;

[0037] Obtain a standard image for each form; construct a standard tensor using the standard images for all forms; the elements of the standard tensor with dimensions i2, j2, and k2 are the elements of the j2-th row and k2-th column of the standard image for the i2-th form;

[0038] Based on the standard tensor and the image tensor corresponding to each tooth, a recognition optimization model corresponding to each tooth is constructed.

[0039] The recognition optimization model corresponding to each tooth is solved to obtain the morphological recognition result information of the tooth;

[0040] By utilizing the morphological recognition results of all teeth, a set of morphological recognition results information is constructed.

[0041] A second aspect of this invention discloses a recognition and processing device based on multi-angle anterior teeth photographs, the device comprising:

[0042] Memory containing executable program code;

[0043] A processor coupled to the memory;

[0044] The processor calls the executable program code stored in the memory to execute the recognition and processing method based on multi-angle anterior teeth photographs.

[0045] In a third aspect of this invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions, which, when invoked by a computer, are used to execute the recognition and processing method based on multi-angle anterior teeth photographs.

[0046] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the recognition and processing method based on multi-angle anterior teeth photographs.

[0047] The beneficial effects of this invention are as follows:

[0048] This invention employs a multi-method fusion strategy, combining image denoising algorithms, image enhancement processing, and a fusion of dental segmentation networks and image template matching recognition, to effectively improve the accuracy of tooth segmentation. Image denoising removes noise interference from images, enhancement highlights tooth features, and fusing multiple segmentation methods leverages their respective advantages to compensate for the shortcomings of a single method. This results in accurate segmentation of each tooth, reliably separating image subsets of each tooth even when teeth are closely spaced or have defects, providing high-quality input data for subsequent morphological recognition.

[0049] This invention utilizes image tensors and standard tensors to construct an optimized recognition model, and solves it using advanced mathematical methods such as generalized eigenvalue decomposition, enabling accurate identification of tooth morphology. This method not only considers the three-dimensional morphological features of teeth but also improves the accuracy and robustness of the recognition results through optimized model construction and solution. Compared with traditional recognition methods based on template matching or simple feature extraction, this invention is better adaptable to different patients, different shooting angles, and various complex tooth morphologies, exhibiting stronger generalization ability and accurately identifying tooth morphological sequence values, providing reliable morphological information for oral diagnosis and treatment.

[0050] This invention acquires multi-angle images of a user's anterior teeth and performs comprehensive processing and analysis on these images. Compared to traditional single-angle image analysis methods, multi-angle images provide more comprehensive tooth morphology information, helping to more accurately assess the three-dimensional morphological features of teeth, such as tooth rotation angles and labial / lingual morphology. This multi-angle information fusion approach further improves the accuracy and reliability of morphological recognition, providing strong support for precise diagnosis in oral medicine. Attached Figure Description

[0051] Figure 1 This is a flowchart illustrating the implementation of the method of the present invention. Detailed Implementation

[0052] To better understand the content of this invention, an embodiment is provided here.

[0053] Figure 1 This is a flowchart illustrating the implementation of the method of the present invention.

[0054] In a first aspect, this invention discloses a method for identifying and processing anterior teeth based on multi-angle anterior teeth photographs, comprising:

[0055] S1, acquire a set of image information of the user's anterior teeth; the set of image information includes multiple subsets of images from different angles; the subset of images from different angles is a set of image information of the user's anterior teeth captured from one angle;

[0056] S2, preprocess the image information set to obtain a set of tooth images; the set of tooth images includes a subset of images for each tooth;

[0057] S3, perform morphological recognition processing on the set of tooth images to obtain a set of morphological recognition result information; the set of morphological recognition result information includes the morphological recognition result information of each tooth;

[0058] The preprocessing of the image information set to obtain a set of tooth images includes:

[0059] S21, perform noise reduction processing on the image information set to obtain a first image set;

[0060] S22, perform image enhancement processing on the first image set to obtain the second image set;

[0061] S23, perform tooth segmentation processing on the second image set to obtain a tooth image set.

[0062] The noise reduction process can be implemented using image noise reduction algorithms.

[0063] The step of performing image enhancement processing on the first image set to obtain the second image set includes:

[0064] For each image in the first image set, a grayscale histogram is calculated to obtain the corresponding grayscale histogram;

[0065] Perform gray-level value statistical processing on the gray-level histogram to obtain the gray-level probability density function;

[0066] The gray-level probability density function is inverted to obtain the inverse gray-level probability density function;

[0067] The gray-level inverse probability density function is subjected to integral enhancement transformation to obtain the enhancement function;

[0068] Using the enhancement function, the pixels of each image in the first image set are calculated to obtain the corresponding enhanced pixels. Using all the enhanced pixels, the corresponding enhanced image is constructed.

[0069] Using all the enhanced images, a second image set is constructed.

[0070] The expression for the integral enhancement transform process is:

[0071]

[0072] Where p(v) represents the inverse probability density function of gray level, v represents the gray level value, q represents the gray level value of the input of the enhancement function, and z(q) represents the enhancement function.

[0073] The expression for the integral enhancement transform processing effectively adjusts the grayscale distribution of an image by applying an integral enhancement transform to the inverse grayscale probability density function. In actual tooth images, due to factors such as uneven lighting, the contrast between teeth and the background or other tissues (such as gums) may be low. This enhancement processing makes the edges and details of the teeth clearer, highlighting the features of the teeth and providing better images for subsequent tooth segmentation and morphological recognition. The inverse grayscale probability density function in the formula considers the distribution of different grayscale values ​​in the image. This grayscale distribution-based enhancement method can adaptively adjust the contrast of the image, achieving a suitable enhancement effect regardless of whether the grayscale distribution of the input image is dark, bright, or has complex grayscale variations. For example, under poor shooting angles or lighting conditions, tooth images may appear generally dark or have overly bright areas. This formula can effectively improve this situation, making the grayscale distribution of the tooth image more uniform, facilitating subsequent recognition processing.

[0074] The step of performing tooth segmentation processing on the second image set to obtain a tooth image set includes:

[0075] S231, using a preset dental segmentation network, perform tooth segmentation processing on the second image set to obtain a first recognition result set; the first recognition result set includes the first sequence number information of teeth in the sub-image of each second image in the second image set;

[0076] S232, perform image template matching and recognition processing on the second image set to obtain a second recognition result set; the second recognition result set includes the second sequence number information of the teeth of each sub-image of the second image in the second image set;

[0077] S233, perform a first fusion calculation on the first recognition result set and the second recognition result set to obtain a set of tooth images.

[0078] The first fusion calculation of the first recognition result set and the second recognition result set to obtain a set of tooth images includes:

[0079] S2331, For each sub-image in the first recognition result set and the second recognition result set, calculate the mean, variance and median value of all pixel gray values ​​of the sub-image respectively;

[0080] S2332, For the same sub-image in the first recognition result set and the second recognition result set, perform pixel value fusion and weighted calculation on the corresponding first tooth number information and second tooth number information to obtain the corresponding tooth number information;

[0081] S2333, using all sub-images corresponding to the same tooth serial number information, construct a subset of images of the teeth corresponding to the tooth serial number information;

[0082] S2334, using all the tooth image subsets, construct a tooth image set.

[0083] The expression for the pixel value fusion weighted calculation is:

[0084]

[0085] Where α1, β1, and ∈1 are the mean, variance, and median of all pixel gray values ​​in the sub-images of the first recognition result set, respectively; α2, β2, and ∈2 are the mean, variance, and median of all pixel gray values ​​in the sub-images of the second recognition result set, respectively; and N1 and N2 are the first and second sequence numbers of the teeth, respectively. and These represent rounding down and rounding up, respectively. N0 represents the tooth number information, and T2() is the second-order Legendre polynomial.

[0086] This formula fuses the first set of recognition results obtained from a dental arch segmentation network and the second set of recognition results obtained from image template matching. This fusion method fully leverages the advantages of both different recognition methods. The dental arch segmentation network can identify the overall structure of teeth based on a deep learning model, while image template matching can match local features using existing tooth templates. The weighted calculation in the formula comprehensively considers the reliability of both recognition results, resulting in more accurate tooth numbering information. For example, in some cases, the dental arch segmentation network may not be accurate enough in identifying local details of teeth, while image template matching can supplement this information; and vice versa. Through fusion calculation, the shortcomings of a single method can be compensated for, improving the accuracy of tooth segmentation.

[0087] The statistical features used in the pixel value fusion weighting calculation reflect the local characteristics of the image, such as brightness, contrast, and texture. By incorporating these statistical features into the fusion calculation, the formula can dynamically adjust the fusion weights according to the actual characteristics of the image. For example, when the variance of a sub-image is large, indicating that the grayscale changes in that image region are complex, the formula will adjust the fusion weights according to this complexity, making the fusion result more consistent with the characteristics of the actual image.

[0088] The morphological recognition processing of the tooth image set to obtain a morphological recognition result information set includes:

[0089] Using the image subset of each tooth in the tooth image set, an image tensor corresponding to the tooth is constructed; the image tensor corresponding to the tooth is a three-dimensional tensor; the elements of the three-dimensional tensor with dimensions i1, j1 and k1 are the gray values ​​of the pixels in the j1th row and k1th column of the i1th sub-image of the tooth image subset;

[0090] Obtain a standard image for each form; construct a standard tensor using the standard images for all forms; the elements of the standard tensor with dimensions i2, j2, and k2 are the elements of the j2-th row and k2-th column of the standard image for the i2-th form;

[0091] Based on the standard tensor and the image tensor corresponding to each tooth, a recognition optimization model corresponding to each tooth is constructed.

[0092] The recognition optimization model corresponding to each tooth is solved to obtain the morphological recognition result information of the tooth;

[0093] By utilizing the morphological recognition results of all teeth, a set of morphological recognition results information is constructed.

[0094] The expression for the recognition optimization model is:

[0095]

[0096] Where M1 and M2 are the row and column dimensions of the sub-image, respectively, N3 is the number of sub-images contained in the image subset, and F j1,k1 Let A(i1,j1,k1) represent the element in the j1-th row and k1-th column of the feature rotation matrix, where A(i1,j1,k1) represents the element of the image tensor corresponding to the tooth with dimensions i1,j1,k1, and B(i2,j1,k1) represent the element of the standard tensor with dimensions i2,j1, andk1. i1,i2 (j1,k1) represents the element in the j1-th row and k1-th column of the generalized feature matrix corresponding to elements i1 and i2, and N4 represents the total number of morphologies. The generalized feature matrix corresponding to elements i1 and i2 is constructed by using a submatrix A(i1,:,:) of the image tensor and a submatrix B(i2,:,:) of the standard tensor to obtain a matrix bundle. The matrix bundle is then subjected to generalized eigenvalue decomposition to obtain generalized eigenvalues ​​and generalized eigenvectors. Each generalized eigenvector is used as a column vector to construct the generalized feature matrix corresponding to elements i1 and i2. exp represents exponential operation, i2 is the vector to be solved, representing the morphology recognition result of the tooth, and is the morphology index value obtained by recognition.

[0097] This formula constructs the image tensor and standard tensor corresponding to the teeth based on three-dimensional tensors. Three-dimensional tensors can fully consider information from multiple dimensions of the tooth image, such as rows, columns, and sub-images, providing a more comprehensive description of tooth morphological features compared to traditional two-dimensional image analysis methods. For example, three-dimensional features such as tooth rotation angle and labial / lingual morphology can be represented through multiple dimensions of the three-dimensional tensor. This comprehensive feature description allows the recognition optimization model to more accurately capture subtle differences in tooth morphology, providing strong support for precise morphological recognition. The feature rotation matrix, through its complex mathematical expression (including exponential, tangent, and sine functions), can rotate and adjust image features to better match the feature space of the standard tensor. The generalized feature matrix is ​​obtained by constructing a matrix bundle from the sub-matrices of the image tensor and standard tensor and performing generalized eigenvalue decomposition. It can extract the essential feature relationship between the image and the standard morphology. This feature extraction and adjustment method allows the model to adapt to differences in tooth morphology. Even in cases where teeth have non-standard morphologies such as wear or rotation, the optimal matching morphology can be found through model optimization, improving the robustness and adaptability of the matching.

[0098] The formula solves for tooth morphology recognition by minimizing the objective function. This optimization method automatically adjusts model parameters to minimize the difference between the tooth image and the standard morphology. During the solution process, the model comprehensively considers the feature information of all sub-images and the similarities and differences between different morphologies, ultimately obtaining the most accurate morphology index value. Compared with traditional recognition methods based on fixed rules or simple matching, this model-based recognition method can better handle complex image features and morphological changes, achieving accurate tooth morphology recognition and providing reliable morphological evidence for oral diagnosis and treatment.

[0099] The solution to the identification optimization model can be obtained using genetic algorithms, ant colony algorithms, Lattice algorithms, etc.

[0100] The morphology refers to the growth and development morphology or healthy morphology of the teeth.

[0101] The preset dentition segmentation network can be the TgNet dentition segmentation network.

[0102] The image template matching and recognition process can adopt a template matching method based on image grayscale, that is, use each tooth template image to match the image to be segmented to obtain the tooth sequence information in the image to be segmented, or adopt a two-stage template matching algorithm.

[0103] The standard images of each morphology can be obtained by using an image sensor to capture the standard morphology.

[0104] The technical solution of this invention is highly adaptable to different shooting conditions and tooth conditions. Through image enhancement processing and optimized segmentation and recognition algorithms, it can still accurately identify tooth morphology under complex conditions such as uneven lighting and noise interference, as well as when teeth have non-standard shapes such as wear and twisting. This strong adaptability makes the method of this invention widely applicable in practical clinical applications, and can meet the needs of anterior tooth morphology recognition in different scenarios.

[0105] A second aspect of this invention discloses a recognition and processing device based on multi-angle anterior teeth photographs, 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 recognition and processing method based on multi-angle anterior teeth photographs.

[0109] In a third aspect of this invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions, which, when invoked by a computer, are used to execute the recognition and processing method based on multi-angle anterior teeth photographs.

[0110] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the recognition and processing method based on multi-angle anterior teeth photographs.

[0111] A fifth aspect of this invention discloses a recognition and processing device based on multi-angle anterior teeth photographs, used to implement the recognition and processing method based on multi-angle anterior teeth photographs as described in the first aspect of the invention, comprising: an image acquisition module and an image recognition module; the image acquisition module is used to acquire a set of image information of a user's anterior teeth; the image information set includes multiple angle image subsets; the angle image subset is a set of image information of a user's anterior teeth taken from one angle; the image recognition module is used to preprocess the image information set to obtain a set of tooth images; and to perform morphological recognition processing on the set of tooth images to obtain a set of morphological recognition result information.

[0112] 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 recognition processing method based on multi-angle front tooth photos, characterized by, include: S1, Acquire a set of image information of the user's anterior teeth; the set of image information includes multiple subsets of images from different angles; The angular image subset is a set of image information of the user's anterior teeth obtained from a single angle. S2, preprocess the image information set to obtain a set of tooth images; The set of tooth images, comprising a subset of images for each tooth, includes: S21, perform noise reduction processing on the image information set to obtain a first image set; S22, perform image enhancement processing on the first image set to obtain a second image set, including: For each image in the first image set, a grayscale histogram is calculated to obtain the corresponding grayscale histogram; Perform gray-level value statistical processing on the gray-level histogram to obtain the gray-level probability density function; The gray-level probability density function is inverted to obtain the inverse gray-level probability density function; The gray-level inverse probability density function is subjected to integral enhancement transformation to obtain the enhancement function; Using the enhancement function, the pixels of each image in the first image set are calculated to obtain the corresponding enhanced pixels. Using all the enhanced pixels, the corresponding enhanced image is constructed. Using all the enhanced images, a second image set is constructed; S23, perform tooth segmentation processing on the second image set to obtain a tooth image set, including: S231, using a preset dental segmentation network, perform tooth segmentation processing on the second image set to obtain a first recognition result set; the first recognition result set includes the first sequence number information of teeth in the sub-image of each second image in the second image set; S232, perform image template matching and recognition processing on the second image set to obtain a second recognition result set; the second recognition result set includes the second sequence number information of the teeth of each sub-image of the second image in the second image set; S233, perform a first fusion calculation on the first recognition result set and the second recognition result set to obtain a set of tooth images, including: S2331, For each sub-image in the first recognition result set and the second recognition result set, calculate the mean, variance and median value of all pixel gray values ​​of the sub-image respectively; S2332, For the same sub-image in the first recognition result set and the second recognition result set, perform pixel value fusion and weighted calculation on the corresponding first tooth number information and second tooth number information to obtain the corresponding tooth number information; S2333, using all sub-images corresponding to the same tooth serial number information, construct a subset of images of the teeth corresponding to the tooth serial number information; S2334, using all the tooth image subsets, construct a tooth image set; The expression for the pixel value fusion weighted calculation is: ， wherein, , and are the mean, variance and median value of all pixel gray values of the sub-image of the first identification result set respectively, , and are the mean, variance and median value of all pixel gray values of the sub-image of the second identification result set respectively, N1 and N2 are the tooth first serial number information and tooth second serial number information respectively, and respectively represent the floor and ceiling, represents the tooth serial number information, is the second order Legendre polynomial; S3, perform morphological recognition processing on the set of tooth images to obtain a set of morphological recognition result information; the set of morphological recognition result information includes the morphological recognition result information of each tooth.

2. The recognition and processing method based on multi-angle anterior teeth photographs as described in claim 1, characterized in that, The morphological recognition processing of the tooth image set to obtain a morphological recognition result information set includes: Using the image subset of each tooth in the tooth image set, an image tensor corresponding to the tooth is constructed; the image tensor corresponding to the tooth is a three-dimensional tensor; the elements of the three-dimensional tensor with dimensions i1, j1 and k1 are the gray values ​​of the pixels in the j1th row and k1th column of the i1th sub-image of the tooth image subset; Obtain a standard image for each form; construct a standard tensor using the standard images for all forms; the elements of the standard tensor with dimensions i2, j2, and k2 are the elements of the j2-th row and k2-th column of the standard image for the i2-th form; Based on the standard tensor and the image tensor corresponding to each tooth, a recognition optimization model corresponding to each tooth is constructed. The recognition optimization model corresponding to each tooth is solved to obtain the morphological recognition result information of the tooth; By utilizing the morphological recognition results of all teeth, a set of morphological recognition results information is constructed.

3. A recognition and processing device based on multi-angle anterior teeth photographs, 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 recognition processing method based on multi-angle anterior teeth photographs as described in any one of claims 1 to 2.

4. 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 recognition and processing method based on multi-angle anterior teeth photographs as described in any one of claims 1 to 2.

5. An information data processing terminal, characterized in that, The information data processing terminal is used to implement the recognition and processing method based on multi-angle anterior teeth photographs as described in any one of claims 1 to 2.

Images