Oral tooth examination method and device

Abstract

The invention discloses an oral cavity tooth examination method and device, and the method comprises the steps: carrying out the preprocessing of a CBCT image of oral cavity teeth; carrying out coarse segmentation on the CBCT image; determining a target segmentation point of the CBCT image after coarse segmentation, and performing fine segmentation processing on the CBCT image after coarse segmentation based on the target segmentation point to obtain a tooth segmentation image; performing color feature extraction processing on the tooth segmentation image based on an octree to obtain tooth color features; calculating the inclination degree of each tooth in the tooth segmentation image, and carrying out tooth crack detection on the tooth segmentation image based on Hough space to obtain a tooth crack detection result; and performing tooth abnormality examination based on the tooth color features, the inclination degree of each tooth and the tooth crack detection result. According to the invention, the obtained tooth abnormality examination result is more comprehensive and accurate, so that the tooth abnormality examination can achieve a more ideal effect.

Description

Technical Field

[0001] The present invention mainly relates to the field of image processing technology, and in particular to an oral teeth inspection method and device. Background Art

[0002] As people's quality of life continues to improve, dental health has gradually attracted people's high attention, and oral dental examinations have also received more and more attention. In the current oral dental examination, the CBCT images of the oral teeth are usually examined and analyzed to determine whether the user's oral teeth are abnormal. In the process of inspection and analysis, the CBCT images of the oral teeth need to be segmented to obtain tooth information. In the current CBCT image segmentation, an adaptive threshold segmentation method is usually used, but this method is prone to blurring of adjacent tooth boundaries and uneven tooth grayscale values, resulting in insufficient reliability of CBCT image segmentation, and thus unable to accurately obtain tooth information. Tooth crack detection is also an important part of oral dental examination. The current tooth crack detection is usually detected by dyeing agents supplemented by image analysis, but this method is prone to oral contamination, and its detection accuracy is also low. At the same time, in the current inspection and analysis of the CBCT images of oral teeth, there is a lack of analysis of the degree of tooth inclination, which makes the obtained tooth inspection results not comprehensive enough, resulting in the failure of the tooth inspection to achieve the expected effect. Summary of the invention

[0003] The purpose of the present invention is to overcome the deficiencies of the prior art. The present invention provides an oral dental examination method and device, so that the obtained dental abnormality examination results are more comprehensive and accurate, so that the dental abnormality examination can achieve a more ideal effect.

[0004] In order to solve the above technical problems, the present invention provides an oral tooth inspection method, the method comprising:

[0005] Acquire a CBCT image of oral teeth, and preprocess the CBCT image to obtain a preprocessed CBCT image;

[0006] Performing rough segmentation on the preprocessed CBCT image to obtain a rough segmented CBCT image;

[0007] Determining target segmentation points of the roughly segmented CBCT image, and performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points to obtain a tooth segmentation image;

[0008] Performing color feature extraction processing on the tooth segmentation image based on the octree to obtain tooth color features;

[0009] Calculating the inclination degree of each tooth in the tooth segmentation image, and performing tooth crack detection on the tooth segmentation image based on the Hough space to obtain a tooth crack detection result;

[0010] A tooth abnormality inspection is performed based on the tooth color characteristics, the inclination degree of each tooth and the tooth crack detection result to obtain a tooth abnormality inspection result.

[0011] Optionally, preprocessing the CBCT image to obtain a preprocessed CBCT image includes:

[0012] Performing format conversion processing on the CBCT image to obtain a CBCT image after format conversion processing;

[0013] Performing grayscale transformation on the CBCT image after the format conversion process to obtain a CBCT image after the grayscale transformation process;

[0014] The CBCT image after grayscale transformation is subjected to filtering, denoising and standardization to obtain the preprocessed CBCT image.

[0015] Optionally, performing coarse segmentation on the preprocessed CBCT image to obtain a coarsely segmented CBCT image includes:

[0016] Calculating the tooth region area of ​​the preprocessed CBCT image, and determining the center point of the upper and lower rows of teeth in the preprocessed CBCT image based on the tooth region area;

[0017] The preprocessed CBCT image is roughly segmented based on the central point to obtain a roughly segmented CBCT image.

[0018] Optionally, determining a target segmentation point of the roughly segmented CBCT image, and performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation point to obtain a tooth segmentation image includes:

[0019] Acquiring cross-sectional control points of the roughly segmented CBCT image, and constructing a coronal segmentation frame based on the cross-sectional control points;

[0020] Determining target segmentation points of the roughly segmented CBCT image based on the coronal segmentation frame;

[0021] Based on the target segmentation points combined with the distance regularized level set evolution model, the roughly segmented CBCT image is finely segmented to obtain a tooth segmentation image.

[0022] Optionally, performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points in combination with a distance regularized level set evolution model to obtain a tooth segmentation image includes:

[0023] Based on the target segmentation point, the watershed algorithm is used to perform threshold segmentation on the roughly segmented CBCT image to obtain a threshold segmented CBCT image;

[0024] The CBCT images after threshold segmentation were iteratively segmented based on the distance regularized level set evolution model to obtain the tooth segmentation image.

[0025] Optionally, the performing color feature extraction processing on the tooth segmentation image based on the octree to obtain the tooth color feature includes:

[0026] Acquire pixel point data of historical tooth images, calculate a coding array based on the pixel point data of historical tooth images, and construct an initial octree node based on the coding array;

[0027] Establishing octree nodes of each level based on the initial octree nodes, and constructing an octree based on the octree nodes of each level;

[0028] The RGB coordinates of each pixel in the tooth segmentation image are obtained, and the color feature extraction process of the tooth segmentation image is performed using the octree based on the RGB coordinates of each pixel to obtain the tooth color feature.

[0029] Optionally, calculating the inclination degree of each tooth in the tooth segmentation image includes:

[0030] Performing tooth recognition on the tooth segmentation image to obtain a semantic perception effect diagram of the teeth;

[0031] Determine the inclination of each tooth based on the semantic perception effect diagram combined with a preset tooth inclination judgment standard;

[0032] Calculating the tooth width of each tooth in the tooth segmentation image, and determining the tooth inclination aggregation based on the inclination of each tooth and the tooth width;

[0033] The degree of inclination of each tooth is determined based on the inclination of each tooth and the tooth inclination aggregation.

[0034] Optionally, performing tooth crack detection on the tooth segmentation image based on the Hough space to obtain a tooth crack detection result includes:

[0035] Binarizing the teeth segmentation image to obtain a number of edge points;

[0036] Obtain the coordinate value of each edge point, and calculate the target value of the edge point in the Hough space based on the coordinate value of each edge point;

[0037] Based on the coordinate value of each edge point, a field function approximation method is used to perform crack prediction simulation to obtain crack prediction simulation data;

[0038] Tooth crack detection is performed based on the crack prediction simulation data and the target value of the Hough space to obtain a tooth crack detection result.

[0039] Optionally, performing a tooth abnormality inspection based on the tooth color characteristics, the degree of inclination of each tooth and the tooth crack detection result to obtain a tooth abnormality inspection result includes:

[0040] The tooth color characteristics, the degree of inclination of each tooth and the tooth crack detection results are compared with the preset tooth template, and it is determined whether there is any abnormality in the tooth based on the comparison result.

[0041] In addition, the present invention also provides an oral tooth inspection device, the device comprising:

[0042] Image preprocessing module: used for acquiring CBCT images, preprocessing the CBCT images, and obtaining preprocessed CBCT images;

[0043] Image coarse segmentation module: used to perform coarse segmentation on the preprocessed CBCT image to obtain a coarsely segmented CBCT image;

[0044] Image fine segmentation module: used to determine the target segmentation points of the roughly segmented CBCT image, and perform fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points to obtain a tooth segmentation image;

[0045] Color feature extraction module: used for performing color feature extraction processing on the tooth segmentation image based on octree to obtain tooth color features;

[0046] The inclination degree and crack detection module is used to calculate the inclination degree of each tooth in the tooth segmentation image, and perform tooth crack detection on the tooth segmentation image based on the Hough space to obtain a tooth crack detection result;

[0047] Teeth abnormality detection module: used to perform tooth abnormality detection based on the tooth color characteristics, the inclination degree of each tooth and the tooth crack detection results to obtain the tooth abnormality detection results.

[0048] In the embodiment of the present invention, the target segmentation point of the CBCT image after rough segmentation is determined, and the CBCT image after rough segmentation is finely segmented based on the target segmentation point, which can improve the reliability of CBCT image segmentation and avoid the problem of blurred adjacent tooth boundaries and uneven tooth grayscale values ​​in the segmented image. The color feature extraction processing of the tooth segmentation image based on the octree can improve the accuracy of tooth color feature extraction and avoid the deviation between the obtained tooth color feature and the actual situation. The inclination degree of each tooth is determined based on the inclination degree of each tooth and the tooth inclination aggregation degree, and the analysis of the tooth inclination degree is added to make the obtained tooth inspection results more comprehensive. The tooth crack detection of the tooth segmentation image based on the Hough space can improve the accuracy and efficiency of tooth crack detection and avoid oral contamination to the user. The tooth abnormality inspection is performed based on the tooth color feature, the inclination degree of each tooth and the tooth crack detection result, so that the obtained tooth abnormality inspection results are more comprehensive and accurate, so that the tooth abnormality inspection can achieve a more ideal effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 is a flow chart of an oral teeth inspection method in an embodiment of the present invention;

[0051] Figure 2 is a flow chart of an oral and dental examination method in another embodiment of the present invention;

[0052] Figure 3 It is a schematic diagram of the structure of the oral teeth inspection device in an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] Embodiment 1

[0055] See also Figure 1 , Figure 1: is a flow chart of an oral tooth inspection method in an embodiment of the present invention, the method comprising:

[0056] S11: Acquire a CBCT image of oral teeth, and preprocess the CBCT image to obtain a preprocessed CBCT image;

[0057] In a specific implementation of the present invention, the preprocessing of the CBCT image to obtain a preprocessed CBCT image includes: performing format conversion processing on the CBCT image to obtain a CBCT image after format conversion processing; performing grayscale transformation processing on the CBCT image after format conversion processing to obtain a CBCT image after grayscale transformation processing; performing filtering, denoising and standardization processing on the CBCT image after grayscale transformation processing to obtain a preprocessed CBCT image.

[0058] Specifically, a cone beam computed tomography (CBCT) image of the user's oral teeth is obtained through a scanning device. The CBCT image is format-converted to obtain a CBCT image after format conversion, and the obtained CBCT image is converted into a file format for computer processing. The CBCT image after format conversion is grayscale transformed, and the grayscale of the CBCT image after format conversion is grayscale transformed by window technology. The window width and window position are used for different simulated grayscale displays. Therefore, the grayscale transformation can be performed by combining the window width and window position through the window technology function to obtain the CBCT image after grayscale transformation. The CBCT image after grayscale transformation is filtered, denoised and standardized. The CBCT image after grayscale transformation is smoothed and filtered by a bilateral filter while maintaining edge features. After filtering and denoising, the CBCT image is standardized, that is, the CBCT image data is de-meaned to achieve centralization. After completing the standardization, the pre-processed CBCT image can be obtained.

[0059] S12: performing rough segmentation on the preprocessed CBCT image to obtain a rough segmented CBCT image;

[0060] In a specific implementation of the present invention, the coarse segmentation of the preprocessed CBCT image to obtain the coarsely segmented CBCT image includes: calculating the tooth region area of ​​the preprocessed CBCT image, and determining the center point of the upper and lower rows of teeth in the preprocessed CBCT image based on the tooth region area; and coarse segmentation of the preprocessed CBCT image based on the center point to obtain the coarsely segmented CBCT image.

[0061] Specifically, the maximum density projection method is used to calculate the tooth area of ​​the preprocessed CBCT image, and the preprocessed CBCT image is projected in the x, y and z directions. The maximum projection value in the three directions is taken to obtain the tooth area, which is the area surrounding the teeth. The center points of the upper and lower rows of teeth in the preprocessed CBCT image are determined based on the tooth area, and the upper and lower rows of teeth in the preprocessed CBCT image are selected according to the tooth area. The corresponding center points are detected according to the selected upper and lower rows of teeth. The preprocessed CBCT image is roughly segmented based on the center point, with the center point as the foreground, and the preset threshold outside the center point as the background. The inside and outside are marked according to the foreground and the background, and then a gradient operation is performed to roughly segment the CBCT image to obtain a roughly segmented CBCT image.

[0062] S13: determining a target segmentation point of the roughly segmented CBCT image, and performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation point to obtain a tooth segmentation image;

[0063] In a specific implementation of the present invention, the method of determining target segmentation points of a roughly segmented CBCT image and performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points to obtain a tooth segmentation image includes: acquiring cross-sectional control points of the roughly segmented CBCT image and constructing a coronal segmentation frame based on the cross-sectional control points; determining target segmentation points of the roughly segmented CBCT image based on the coronal segmentation frame; and performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points in combination with a distance regularized level set evolution model to obtain a tooth segmentation image.

[0064] Furthermore, the method performs fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points in combination with the distance regularized level set evolution model to obtain a tooth segmentation image, including: performing threshold segmentation on the roughly segmented CBCT image using a watershed algorithm based on the target segmentation points to obtain a threshold segmented CBCT image; and performing iterative segmentation processing on the threshold segmented CBCT image based on the distance regularized level set evolution model to obtain a tooth segmentation image.

[0065] Specifically, after the rough segmentation, the tooth information in the obtained CBCT image is not prominent enough, so the CBCT image needs to be further finely segmented, and the cross-sectional control points of the CBCT image after the rough segmentation are obtained. The near-lingual control point closest to the lingual side and the near-buccal control point closest to the buccal side are determined in the CBCT image after the rough segmentation, and they are used as cross-sectional control points. A coronal segmentation frame is constructed based on the cross-sectional control points, and the corresponding vector direction is determined according to the cross-sectional control points. A segmentation body is constructed according to the corresponding vector direction and the cross-sectional control points, and the segmentation body is projected to the coronal plane of the CBCT image after the rough segmentation to obtain the coronal segmentation frame. Based on the coronal segmentation frame, the target segmentation points of the CBCT image after the rough segmentation are determined, and the end points of the coronal segmentation frame are obtained. The target segmentation points of the CBCT image after the rough segmentation are constructed according to the end points of the coronal segmentation frame and the cross-sectional control points. Based on the target segmentation point, the watershed algorithm is used to perform threshold segmentation on the CBCT image after rough segmentation. The watershed algorithm is an image processing algorithm based on regional segmentation. By simulating the process of rainwater flowing on the terrain, the watershed is found to determine the segmentation boundary in the image. According to the target segmentation point, the CBCT image after rough segmentation is initialized with double level sets in the alternating order of teeth. The level set is set by the level set function to obtain the initialized CBCT image. The target segmentation point is used as the watershed for threshold cutting to obtain the CBCT image after threshold segmentation. The CBCT images after threshold segmentation are iteratively segmented based on the distance regularized level set evolution model. The distance regularized level set evolution model is a mathematical model for image segmentation. The distance regularized level set evolution model is used to improve the shortcomings of the traditional level set method by introducing the distance regularization term, thereby improving the accuracy and stability of image segmentation. The CBCT images after threshold segmentation are iteratively segmented layer by layer through the distance regularized level set evolution model. In each iterative segmentation process, the double level sets are intersected and merged to avoid the situation where the level set curves are intertwined or merged into one curve during the evolution process. The level set function is solved in an iterative manner to obtain the segmentation result. At the same time, in each iterative segmentation, the grayscale distribution inside the iterative segmentation result curve is Gaussian fitted to improve the segmentation accuracy and reduce the interference of the surrounding area. When the iterative segmentation reaches the preset number of iterations, the tooth segmentation image can be obtained.

[0066] S14: performing color feature extraction processing on the tooth segmentation image based on the octree to obtain tooth color features;

[0067] In the specific implementation process of the present invention, the color feature extraction processing of the tooth segmentation image based on the octree to obtain the tooth color feature includes: obtaining pixel point data of the historical tooth image, calculating the coding array based on the pixel point data of the historical tooth image, and constructing the initial octree node based on the coding array; establishing octree nodes of each level based on the initial octree node, and constructing an octree based on the octree nodes of each level; obtaining the RGB coordinates of each pixel point in the tooth segmentation image, and using the octree to extract the color feature of the tooth segmentation image based on the RGB coordinates of each pixel point to obtain the tooth color feature.

[0068] Specifically, the pixel data of the historical tooth image is obtained from the database; the coding array is calculated based on the pixel data of the historical tooth image, and the calculation space is divided according to the pixel data of the historical tooth image using a preset coding strategy to obtain the corresponding coding value, wherein the preset coding strategy is to use a marking function for encoding, generate a coding array according to the corresponding coding value, and construct an initial octree node based on the coding array, create an octree node array according to the coding array, and configure the key value corresponding to the octree node array, and construct the initial octree node according to the octree node array and the corresponding key value. Based on the initial octree node, establish octree nodes of each level, calculate the number of octree nodes in each level according to the initial octree node, and create a level array and a node address array, the level array is used to record the number of nodes, the node address array is used to record the starting position of the node, generate a tree node array according to the level array, and construct an octree according to the tree node array and the node address array, so that the nodes in the created octree record the corresponding tooth color information. The red, green, blue (RGB) coordinates of each pixel in the tooth segmentation image are obtained, where the RGB coordinates are the RGB channel values ​​of each pixel. The octree is used to perform color feature extraction on the tooth segmentation image based on the RGB coordinates of each pixel. The RGB coordinates of each pixel are sequentially input into the octree. The RGB value of each node in the octree is obtained according to the RGB coordinates of each pixel. The average is calculated according to the RGB values ​​of each node. The tooth color information, that is, the tooth color feature, is obtained according to the average.

[0069] S15: calculating the inclination degree of each tooth in the tooth segmentation image, and performing tooth crack detection on the tooth segmentation image based on the Hough space to obtain a tooth crack detection result;

[0070] In the specific implementation process of the present invention, the calculation of the inclination degree of each tooth in the tooth segmentation image includes: performing tooth recognition on the tooth segmentation image to obtain a semantic perception rendering of the teeth; determining the inclination of each tooth based on the semantic perception rendering combined with a preset tooth inclination judgment standard; calculating the tooth width of each tooth in the tooth segmentation image, and determining the tooth inclination aggregation based on the inclination of each tooth and the tooth width; determining the inclination degree of each tooth based on the inclination of each tooth and the tooth inclination aggregation.

[0071] Furthermore, the tooth crack detection is performed on the tooth segmentation image based on the Hough space to obtain the tooth crack detection result, including: performing edge binarization on the tooth segmentation image to obtain a plurality of edge points; obtaining the coordinate value of each edge point, and calculating the target value of the edge point in the Hough space based on the coordinate value of each edge point; performing crack prediction simulation based on the coordinate value of each edge point using the field function approximation method to obtain crack prediction simulation data; performing tooth crack detection based on the crack prediction simulation data and the target value of the Hough space to obtain the tooth crack detection result.

[0072] Specifically, teeth recognition is performed on the teeth segmentation image, and teeth recognition is performed on the teeth segmentation image through a semantic segmentation network, which is a convergence model obtained by inputting a sample data set into a deep neural network model for training, to obtain a semantic perception effect diagram of teeth. The inclination of each tooth is determined based on the semantic perception effect diagram combined with a preset tooth inclination judgment standard, the preset tooth inclination judgment standard is the inclination judgment standard of teeth in different positions, the root midpoint and the crown midpoint are obtained according to the semantic perception effect diagram of the teeth, the vector of the root midpoint and the crown midpoint is used as the direction of the tooth, and the inclination of the tooth is analyzed according to the direction of the tooth combined with the preset tooth inclination judgment standard. The tooth width of each tooth in the tooth segmentation image is calculated, and the distance between the two edge endpoints of each tooth is determined according to the semantic perception effect diagram, and then the tooth width of each tooth is calculated. The tooth inclination aggregation is determined based on the inclination and tooth width of each tooth. The corresponding inclination level is determined according to the inclination and tooth width of each tooth. The inclination level greater than or equal to the preset inclination level threshold is taken as the abnormal inclination level. When the number of abnormal inclination levels that appear continuously in the tooth segmentation image is greater than or equal to the preset number, the abnormal inclination level that appears continuously is taken as an abnormal subsequence, the number and length of the abnormal subsequence are counted, and the longest abnormal subsequence length is obtained. The tooth inclination aggregation is calculated according to the number of abnormal subsequences and the longest abnormal subsequence length. The calculation expression of the tooth inclination aggregation is:

[0073]

[0074] Among them, F is the degree of tooth inclination aggregation, S is the number of abnormal subsequences, lmax is the longest abnormal subsequence length, and σ is the correction coefficient. The degree of inclination of each tooth is determined based on the inclination of each tooth and the inclination clustering of the teeth, that is, the degree of inclination of each tooth is determined according to the inclination of each tooth and the inclination clustering of the teeth and the corresponding weights. The tooth segmentation image is edge binarized, the coordinates of each pixel point in the tooth segmentation image are obtained, the gradient value of the pixel point is calculated according to the coordinates of each pixel point, the gradient value of the pixel point is compared with the preset gradient threshold, if the gradient value of the pixel point is greater than the preset gradient threshold, then the pixel point is an edge point, and a number of edge points are obtained. The coordinate value of each edge point is obtained, and the target value of the edge point in the Hough space is calculated based on the coordinate value of each edge point, and the target value of the edge point in the Hough space is calculated using the Hough space calculation formula according to the coordinate value of each edge point, that is, the polar radius of the edge point in the Hough space is calculated. Based on the coordinate value of each edge point, the field function approximation method is used to perform crack prediction simulation. The coordinate value of each pixel point is input into the tooth strain simulation software to construct a related three-dimensional model. The three-dimensional model is simulated to apply pressure, that is, the pressure applied to the tooth is simulated to obtain the strain distribution of the tooth. According to the strain distribution of the tooth, the field quantity function is used to predict the development of tooth cracks, and crack prediction simulation data is obtained, that is, the predicted development value of tooth cracks is obtained. Tooth crack detection is performed based on the crack prediction simulation data and the target value of the Hough space. The tooth crack point is determined by the Moore voting method using the target value of the Hough space. The tooth crack level is determined according to the tooth crack point and the crack prediction simulation data, that is, the tooth crack detection result is obtained.

[0075] S16: Perform a tooth abnormality inspection based on the tooth color characteristics, the degree of inclination of each tooth and the tooth crack detection result to obtain a tooth abnormality inspection result.

[0076] In the specific implementation process of the present invention, the tooth abnormality inspection is performed based on the tooth color characteristics, the inclination degree of each tooth and the tooth crack detection results to obtain the tooth abnormality inspection results, including: comparing the tooth color characteristics, the inclination degree of each tooth and the tooth crack detection results with the preset tooth template, and judging whether there is any abnormality in the teeth based on the comparison results.

[0077] Specifically, the tooth color characteristics, the inclination degree of each tooth and the tooth crack detection results are compared with a preset tooth template, which contains normal tooth conditions and an allowable deviation range. That is, the tooth color characteristics of the user's teeth, the inclination degree of each tooth and the tooth crack conditions are compared to see if the deviation from the normal conditions is within the allowable deviation range. If they exceed the allowable deviation range, it means that there is an abnormality in the user's teeth, and subsequent dental treatment is assisted according to the magnitude of the deviation.

[0078] In the embodiment of the present invention, the target segmentation point of the CBCT image after rough segmentation is determined, and the CBCT image after rough segmentation is finely segmented based on the target segmentation point, which can improve the reliability of CBCT image segmentation and avoid the problem of blurred adjacent tooth boundaries and uneven tooth grayscale values ​​in the segmented image. The color feature extraction processing of the tooth segmentation image based on the octree can improve the accuracy of tooth color feature extraction and avoid the deviation between the obtained tooth color feature and the actual situation. The inclination degree of each tooth is determined based on the inclination degree of each tooth and the tooth inclination aggregation degree, and the analysis of the tooth inclination degree is added to make the obtained tooth inspection results more comprehensive. The tooth crack detection of the tooth segmentation image based on the Hough space can improve the accuracy and efficiency of tooth crack detection and avoid oral contamination to the user. The tooth abnormality inspection is performed based on the tooth color feature, the inclination degree of each tooth and the tooth crack detection result, so that the obtained tooth abnormality inspection results are more comprehensive and accurate, so that the tooth abnormality inspection can achieve a more ideal effect.

[0079] Embodiment 2

[0080] See also Figure 2 , Figure 2 FIG. 1 is a flow chart of an oral tooth inspection method according to another embodiment of the present invention, wherein the method comprises:

[0081] S201: Acquire a CBCT image of oral teeth, and preprocess the CBCT image to obtain a preprocessed CBCT image;

[0082] S202: performing rough segmentation on the preprocessed CBCT image to obtain a rough segmented CBCT image;

[0083] S203: Acquire cross-sectional control points of the roughly segmented CBCT image, and construct a coronal segmentation frame based on the cross-sectional control points;

[0084] In a specific implementation process of the present invention, the cross-sectional control points of the roughly segmented CBCT image are obtained, the near-lingual control point closest to the lingual side and the near-buccal control point closest to the buccal side are determined in the roughly segmented CBCT image, and are used as cross-sectional control points. A coronal segmentation frame is constructed based on the cross-sectional control points, and the corresponding vector directions are determined according to the cross-sectional control points. A segmentation body is constructed according to the corresponding vector directions and the cross-sectional control points, and the segmentation body is projected onto the coronal plane of the roughly segmented CBCT image to obtain the coronal segmentation frame.

[0085] S204: determining target segmentation points of the roughly segmented CBCT image based on the coronal segmentation frame;

[0086] In a specific implementation of the present invention, target segmentation points of the roughly segmented CBCT image are determined based on the coronal segmentation frame, endpoints of the coronal segmentation frame are obtained, and target segmentation points of the roughly segmented CBCT image are constructed based on the endpoints of the coronary segmentation frame and cross-sectional control points.

[0087] S205: performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points combined with the distance regularized level set evolution model to obtain a tooth segmentation image;

[0088] In the specific implementation process of the present invention, the watershed algorithm is used to perform threshold segmentation on the roughly segmented CBCT image based on the target segmentation point. The watershed algorithm is an image processing algorithm based on regional segmentation. By simulating the process of rainwater flowing on the terrain, the watershed is found to determine the segmentation boundary in the image. The roughly segmented CBCT image is initialized with double level sets according to the alternating order of teeth according to the target segmentation point. The level set is set by the level set function to obtain the initialized CBCT image. The target segmentation point is used as the watershed for threshold cutting to obtain the CBCT image after threshold segmentation. The CBCT images after threshold segmentation are iteratively segmented based on the distance regularized level set evolution model. The distance regularized level set evolution model is a mathematical model for image segmentation. The distance regularized level set evolution model is used to improve the shortcomings of the traditional level set method by introducing the distance regularization term, thereby improving the accuracy and stability of image segmentation. The CBCT images after threshold segmentation are iteratively segmented layer by layer through the distance regularized level set evolution model. In each iterative segmentation process, the double level sets are intersected and merged to avoid the situation where the level set curves are intertwined or merged into one curve during the evolution process. The level set function is solved in an iterative manner to obtain the segmentation result. At the same time, in each iterative segmentation, the grayscale distribution inside the iterative segmentation result curve is Gaussian fitted to improve the segmentation accuracy and reduce the interference of the surrounding area. When the iterative segmentation reaches the preset number of iterations, the tooth segmentation image can be obtained.

[0089] S206: performing color feature extraction processing on the tooth segmentation image based on the octree to obtain tooth color features;

[0090] S207: Calculating the inclination degree of each tooth in the tooth segmentation image;

[0091] S208: performing edge binarization on the tooth segmentation image to obtain a plurality of edge points;

[0092] In the specific implementation process of the present invention, the edge binarization is performed on the tooth segmentation image, the coordinates of each pixel point in the tooth segmentation image are obtained, the gradient value of the pixel point is calculated according to the coordinates of each pixel point, and the gradient value of the pixel point is compared with a preset gradient threshold. If the gradient value of the pixel point is greater than the preset gradient threshold, the pixel point is an edge point, and a number of edge points are obtained.

[0093] S209: Obtain the coordinate value of each edge point, and calculate the target value of the edge point in the Hough space based on the coordinate value of each edge point;

[0094] In the specific implementation process of the present invention, the coordinate value of each edge point is obtained, and the target value of the edge point in the Hough space is calculated based on the coordinate value of each edge point. The target value of the edge point in the Hough space is calculated using the Hough space calculation formula according to the coordinate value of each edge point, that is, the polar radius of the edge point in the Hough space is calculated.

[0095] S210: performing crack prediction simulation based on the coordinate value of each edge point using a field function approximation method to obtain crack prediction simulation data;

[0096] In the specific implementation process of the present invention, the field function approximation method is used to perform crack prediction simulation based on the coordinate value of each edge point, and the coordinate value of each pixel point is input into the tooth strain simulation software to construct a related three-dimensional model. The three-dimensional model is simulated by applying pressure, that is, the pressure applied to the teeth is simulated to obtain the strain distribution of the teeth, and the field quantity function is used to predict the development of tooth cracks based on the strain distribution of the teeth to obtain crack prediction simulation data, that is, the predicted value of the development of tooth cracks is obtained.

[0097] S211: performing tooth crack detection based on the crack prediction simulation data and the target value of the Hough space to obtain a tooth crack detection result;

[0098] In the specific implementation process of the present invention, tooth crack detection is performed based on the crack prediction simulation data and the target value of the Hough space, the tooth crack point is determined by the target value of the Hough space through the Moore voting method, and the tooth crack level is determined according to the tooth crack point and the crack prediction simulation data, so as to obtain the tooth crack detection result.

[0099] S212: Perform a tooth abnormality inspection based on the tooth color characteristics, the degree of inclination of each tooth and the tooth crack detection result to obtain a tooth abnormality inspection result.

[0100] In the embodiment of the present invention, the target segmentation point of the CBCT image after rough segmentation is determined, and the CBCT image after rough segmentation is finely segmented based on the target segmentation point, which can improve the reliability of CBCT image segmentation and avoid the problem of blurred adjacent tooth boundaries and uneven tooth grayscale values ​​in the segmented image. The color feature extraction processing of the tooth segmentation image based on the octree can improve the accuracy of tooth color feature extraction and avoid the deviation between the obtained tooth color feature and the actual situation. The inclination degree of each tooth is determined based on the inclination degree of each tooth and the tooth inclination aggregation degree, and the analysis of the tooth inclination degree is added to make the obtained tooth inspection results more comprehensive. The tooth crack detection of the tooth segmentation image based on the Hough space can improve the accuracy and efficiency of tooth crack detection and avoid oral contamination to the user. The tooth abnormality inspection is performed based on the tooth color feature, the inclination degree of each tooth and the tooth crack detection result, so that the obtained tooth abnormality inspection results are more comprehensive and accurate, so that the tooth abnormality inspection can achieve a more ideal effect.

[0101] Embodiment 3

[0102] See also Figure 3 , Figure 3 : is a schematic diagram of the structure of an oral tooth inspection device in an embodiment of the present invention, the device comprises:

[0103] Image preprocessing module 31: used for acquiring CBCT images, preprocessing the CBCT images, and obtaining preprocessed CBCT images;

[0104] The image rough segmentation module 32 is used to perform rough segmentation on the preprocessed CBCT image to obtain a rough segmented CBCT image;

[0105] The image fine segmentation module 33 is used to determine the target segmentation points of the roughly segmented CBCT image, and perform fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points to obtain a tooth segmentation image;

[0106] Color feature extraction module 34: used for performing color feature extraction processing on the tooth segmentation image based on octree to obtain tooth color features;

[0107] The inclination degree and crack detection module 35 is used to calculate the inclination degree of each tooth in the tooth segmentation image, and perform tooth crack detection on the tooth segmentation image based on the Hough space to obtain a tooth crack detection result;

[0108] The tooth abnormality detection module 36 is used to perform a tooth abnormality detection based on the tooth color characteristics, the inclination degree of each tooth and the tooth crack detection result to obtain a tooth abnormality detection result.

[0109] In the specific implementation process of the present invention, the specific implementation method of the device item can refer to the implementation method of the above-mentioned method item, which will not be repeated here.

[0110] In the embodiment of the present invention, the target segmentation point of the CBCT image after rough segmentation is determined, and the CBCT image after rough segmentation is finely segmented based on the target segmentation point, which can improve the reliability of CBCT image segmentation and avoid the problem of blurred adjacent tooth boundaries and uneven tooth grayscale values ​​in the segmented image. The color feature extraction processing of the tooth segmentation image based on the octree can improve the accuracy of tooth color feature extraction and avoid the deviation between the obtained tooth color feature and the actual situation. The inclination degree of each tooth is determined based on the inclination degree of each tooth and the tooth inclination aggregation degree, and the analysis of the tooth inclination degree is added to make the obtained tooth inspection results more comprehensive. The tooth crack detection of the tooth segmentation image based on the Hough space can improve the accuracy and efficiency of tooth crack detection and avoid oral contamination to the user. The tooth abnormality inspection is performed based on the tooth color feature, the inclination degree of each tooth and the tooth crack detection result, so that the obtained tooth abnormality inspection results are more comprehensive and accurate, so that the tooth abnormality inspection can achieve a more ideal effect.

[0111] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0112] In addition, the above is a detailed introduction to an oral tooth inspection method and device provided by an embodiment of the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for a general technician in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for oral tooth inspection, characterized in that: The method comprises: Acquire a cone beam computed tomography (CBCT) image of oral teeth, and preprocess the CBCT image to obtain a preprocessed CBCT image; Performing rough segmentation on the preprocessed CBCT image to obtain a rough segmented CBCT image; Determining target segmentation points of the roughly segmented CBCT image, and performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points to obtain a tooth segmentation image; Performing color feature extraction processing on the tooth segmentation image based on the octree to obtain tooth color features; Calculating the inclination degree of each tooth in the tooth segmentation image, and performing tooth crack detection on the tooth segmentation image based on the Hough space to obtain a tooth crack detection result; A tooth abnormality inspection is performed based on the tooth color characteristics, the inclination degree of each tooth and the tooth crack detection result to obtain a tooth abnormality inspection result.

2. The oral tooth inspection method according to claim 1, characterized in that: The preprocessing of the CBCT image to obtain a preprocessed CBCT image includes: Performing format conversion processing on the CBCT image to obtain a CBCT image after format conversion processing; Performing grayscale transformation on the CBCT image after the format conversion process to obtain a CBCT image after the grayscale transformation process; The CBCT image after grayscale transformation is subjected to filtering, denoising and standardization to obtain the preprocessed CBCT image.

3. The oral tooth inspection method according to claim 1, characterized in that: The method of performing rough segmentation on the preprocessed CBCT image to obtain the rough segmented CBCT image includes: Calculating the tooth region area of ​​the preprocessed CBCT image, and determining the center point of the upper and lower rows of teeth in the preprocessed CBCT image based on the tooth region area; The preprocessed CBCT image is roughly segmented based on the central point to obtain a roughly segmented CBCT image.

4. The oral tooth inspection method according to claim 1, characterized in that: The step of determining a target segmentation point of the roughly segmented CBCT image and performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation point to obtain a tooth segmentation image includes: Acquiring cross-sectional control points of the roughly segmented CBCT image, and constructing a coronal segmentation frame based on the cross-sectional control points; Determining target segmentation points of the roughly segmented CBCT image based on the coronal segmentation frame; Based on the target segmentation points combined with the distance regularized level set evolution model, the roughly segmented CBCT image is finely segmented to obtain a tooth segmentation image.

5. The oral tooth inspection method according to claim 4, characterized in that: The method of performing fine segmentation processing on the roughly segmented CBCT image based on the target segmentation point combined with the distance regularized level set evolution model to obtain a tooth segmentation image includes: Based on the target segmentation point, the watershed algorithm is used to perform threshold segmentation on the roughly segmented CBCT image to obtain a threshold segmented CBCT image; The CBCT images after threshold segmentation were iteratively segmented based on the distance regularized level set evolution model to obtain the tooth segmentation image.

6. The oral tooth inspection method according to claim 1, characterized in that: The color feature extraction process is performed on the tooth segmentation image based on the octree to obtain the tooth color feature, including: Acquire pixel point data of historical tooth images, calculate a coding array based on the pixel point data of historical tooth images, and construct an initial octree node based on the coding array; Establishing octree nodes of each level based on the initial octree nodes, and constructing an octree based on the octree nodes of each level; The red, green, and blue (RGB) coordinates of each pixel in the tooth segmentation image are obtained, and the color feature extraction process of the tooth segmentation image is performed using the octree based on the RGB coordinates of each pixel to obtain the tooth color feature.

7. The oral tooth inspection method according to claim 1, characterized in that: The calculating the inclination degree of each tooth in the tooth segmentation image comprises: Performing tooth recognition on the tooth segmentation image to obtain a semantic perception effect diagram of the teeth; Determine the inclination of each tooth based on the semantic perception effect diagram combined with a preset tooth inclination judgment standard; Calculating the tooth width of each tooth in the tooth segmentation image, and determining the tooth inclination aggregation based on the inclination of each tooth and the tooth width; The degree of inclination of each tooth is determined based on the inclination of each tooth and the tooth inclination aggregation.

8. The oral tooth inspection method according to claim 1, characterized in that: The step of performing tooth crack detection on the tooth segmentation image based on the Hough space to obtain a tooth crack detection result includes: Binarizing the teeth segmentation image to obtain a number of edge points; Obtain the coordinate value of each edge point, and calculate the target value of the edge point in the Hough space based on the coordinate value of each edge point; Based on the coordinate value of each edge point, a field function approximation method is used to perform crack prediction simulation to obtain crack prediction simulation data; Tooth crack detection is performed based on the crack prediction simulation data and the target value of the Hough space to obtain a tooth crack detection result.

9. The oral tooth inspection method according to claim 1, characterized in that: The tooth abnormality inspection is performed based on the tooth color characteristics, the inclination degree of each tooth and the tooth crack detection result to obtain the tooth abnormality inspection result, including: The tooth color characteristics, the degree of inclination of each tooth and the tooth crack detection results are compared with the preset tooth template, and it is determined whether there is any abnormality in the tooth based on the comparison result.

10. An oral tooth inspection device, characterized in that: The device comprises: Image preprocessing module: used for acquiring CBCT images, preprocessing the CBCT images, and obtaining preprocessed CBCT images; Image coarse segmentation module: used to perform coarse segmentation on the preprocessed CBCT image to obtain a coarsely segmented CBCT image; Image fine segmentation module: used to determine the target segmentation points of the roughly segmented CBCT image, and perform fine segmentation processing on the roughly segmented CBCT image based on the target segmentation points to obtain a tooth segmentation image; Color feature extraction module: used for performing color feature extraction processing on the tooth segmentation image based on octree to obtain tooth color features; The inclination degree and crack detection module is used to calculate the inclination degree of each tooth in the tooth segmentation image, and perform tooth crack detection on the tooth segmentation image based on the Hough space to obtain a tooth crack detection result; Teeth abnormality detection module: used to perform tooth abnormality detection based on the tooth color characteristics, the inclination degree of each tooth and the tooth crack detection results to obtain the tooth abnormality detection results.

Images