Impacted tooth auxiliary identification method based on oral three-dimensional image

Abstract

This invention relates to the field of oral image analysis technology, specifically to a method for assisting in the identification of impacted teeth based on three-dimensional oral images. The method filters edge pixels in oral CT images to obtain enhanced images based on the degree of grayscale variation of these pixels and the shape, edge clarity, and grayscale distribution of their respective analysis areas. Then, based on the shape and tilt differences between the enhanced sagittal CT image and other tooth regions of the wisdom tooth in the three-dimensional oral model, the area of ​​the corresponding tooth region in the enhanced axial CT image, and the distance between adjacent tooth regions in the enhanced coronal CT image, the impacted wisdom tooth in the three-dimensional oral model is determined. This invention effectively improves the accuracy of impacted wisdom tooth identification.

Description

Technical Field

[0001] This invention relates to the field of oral image analysis technology, specifically to a method for assisting in the identification of impacted teeth based on three-dimensional oral images. Background Technology

[0002] Teeth that are improperly positioned within the jawbone and cannot erupt into their normal occlusal position are called impacted teeth or impacted wisdom teeth. The most common impacted tooth is the mandibular third molar, also known as an impacted wisdom tooth. Impacted teeth and the gums covering them easily trap food debris, breed bacteria, and cause bad breath. When the body's resistance is lowered, inflammation often occurs. Therefore, it is necessary to identify impacted teeth in the mouth for timely treatment.

[0003] In the process of identifying impacted wisdom teeth, X-ray images of the oral cavity are usually acquired from a certain perspective. The impacted wisdom teeth are identified based on features such as the position, angle, and eruption status of the wisdom teeth in the image. However, in images from a single perspective, there may be situations where teeth obscure the wisdom teeth, which can cause errors in the relevant data of the teeth in the image, resulting in a low accuracy rate for identifying impacted wisdom teeth. Summary of the Invention

[0004] To address the technical problem of low accuracy in identifying impacted wisdom teeth due to errors in oral images from a single perspective, this invention aims to provide an auxiliary method for identifying impacted wisdom teeth based on three-dimensional oral imaging. The specific technical solution adopted is as follows:

[0005] This invention proposes a method for assisted identification of impacted teeth based on three-dimensional oral imaging, the method comprising:

[0006] Acquire oral CT images of the patient, including: oral axial CT images, oral coronal CT images and oral sagittal CT images;

[0007] The analysis region in each oral CT image is obtained; based on the gray level change of each edge pixel in each oral CT image, the shape, edge clarity and gray level distribution of the analysis region where each edge pixel is located, the edge pixels in each oral CT image are filtered to obtain the enhanced image of each oral CT image.

[0008] The tooth regions in the enhanced images of each oral CT image are obtained; the enhanced images of all oral CT images are reconstructed in three dimensions to obtain a three-dimensional oral model, and the wisdom teeth in the three-dimensional oral model are identified; based on the shape and tilt differences between the corresponding tooth region of each wisdom tooth in the enhanced images of the sagittal CT images of the oral model and the other tooth regions, the area of ​​the corresponding tooth region of each wisdom tooth in the enhanced images of the axial CT images of the oral model, and the distance between adjacent tooth regions in the enhanced images of the coronal CT images of the oral model, the impacted wisdom teeth in the patient's three-dimensional oral model are determined.

[0009] Furthermore, the method for acquiring the enhanced image of each oral CT image includes:

[0010] Based on the degree of grayscale change of each edge pixel in each oral CT image, the shape, edge clarity and grayscale distribution of the analysis area where each edge pixel is located, the possible tooth values ​​of each edge pixel in each oral CT image are obtained.

[0011] For each oral CT image, one edge pixel in the oral CT image is randomly selected and recorded as the point to be filtered. The possible values ​​of the tooth are used as the weights of the gray values ​​of the remaining edge pixels excluding the point to be filtered within the preset filtering window of the point to be filtered. The weighted average value is then used as the filtered gray value of the point to be filtered.

[0012] The gray values ​​of edge pixels in the oral CT image are updated using the filtered gray values ​​to obtain an enhanced image of the oral CT image.

[0013] Furthermore, obtaining the possible tooth values ​​for each edge pixel in each oral CT image includes:

[0014] For each oral CT image, the shape grayscale index of each analysis region is obtained based on the shape and grayscale distribution of each analysis region in the oral CT image;

[0015] The mean of the absolute values ​​of the differences in gradient values ​​between two adjacent edge pixels on the edge of each analysis region in an oral CT image is calculated as the edge clarity index for each analysis region.

[0016] Based on the gradient value of each edge pixel in the oral CT image, as well as the shape grayscale index and edge clarity index of the analysis region where each edge pixel is located, the possible tooth values ​​of each edge pixel in the oral CT image are obtained.

[0017] Further, obtaining the shape grayscale index of each analysis region includes:

[0018] For each analysis region in each oral CT image, obtain the discrete index of the gray values ​​of all pixels in the analysis region; calculate the central value of the gray values ​​of all pixels in the analysis region.

[0019] Obtain the geometric centroid and grayscale centroid of the analysis region, and use the distance between the geometric centroid and the grayscale centroid as the shape rule value;

[0020] Based on the discrete index, the setpoint, and the shape rule value, the shape grayscale index of the analysis area is obtained; the discrete index and the shape rule value are both negatively correlated with the shape grayscale index, and the setpoint is positively correlated with the shape grayscale index.

[0021] Further, the determination of impacted wisdom teeth in the patient's three-dimensional oral model based on the shape and tilt differences between the corresponding tooth region and other tooth regions in the enhanced sagittal CT image of each wisdom tooth in the three-dimensional oral model, the area of ​​the corresponding tooth region in the enhanced axial CT image of each wisdom tooth in the three-dimensional oral model, and the distance between adjacent tooth regions in the enhanced coronal CT image of the three-dimensional oral model includes:

[0022] Obtain the control tooth for each wisdom tooth in the patient's three-dimensional oral model; based on the shape difference and tilt difference of the corresponding tooth region in the enhanced image of the sagittal CT image of the oral cavity, obtain the growth deviation value of each wisdom tooth.

[0023] The distance between two adjacent tooth regions in the enhanced image of the coronal CT image of the oral cavity is obtained; the tooth distance set of each wisdom tooth is constructed by the distance between the two adjacent tooth regions in the enhanced image of the coronal CT image of each control tooth in the oral cavity 3D model; the sum of the absolute values ​​of the differences between the distance between the two adjacent tooth regions in the enhanced image of the coronal CT image of each wisdom tooth in the oral cavity 3D model and each element in the tooth distance set is used as the normal distance difference value of each wisdom tooth.

[0024] Based on the area of ​​the corresponding tooth region in the enhanced image of the axial CT image of each wisdom tooth in the three-dimensional oral model, the difference between the growth deviation value and the normal interval value, the impaction characteristic value of each wisdom tooth is obtained; the area and the impaction characteristic value are negatively correlated, and the growth deviation value and the difference between the normal interval value are both positively correlated with the impaction characteristic value.

[0025] The impacted wisdom teeth in the patient's three-dimensional oral cavity model were determined using the impacted characteristic values.

[0026] Furthermore, obtaining the growth deviation value of each wisdom tooth includes:

[0027] Obtain the minimum bounding rectangle of each tooth region in the enhanced image of the sagittal CT image of the oral cavity;

[0028] The angle between the direction of the long side of the minimum bounding rectangle of each tooth region and the preset direction is recorded as the direction index of the corresponding tooth region; the absolute value of the difference between the length and width of the minimum bounding rectangle of each tooth region is calculated as the shape index of each tooth region.

[0029] Based on the differences in the directional and shape indices of the corresponding tooth regions in the enhanced sagittal CT images of the patient's oral 3D model, the local normal deviation value of each wisdom tooth and each control tooth is obtained; the sum of the local growth deviation values ​​of each wisdom tooth and all control teeth in the oral 3D model is taken as the growth deviation value of each wisdom tooth.

[0030] Further, the step of using the impacted characteristic value to determine the impacted wisdom teeth among all wisdom teeth in the patient's three-dimensional oral model includes:

[0031] For all wisdom teeth in the patient's oral cavity 3D model, the wisdom teeth with impacted characteristic values ​​greater than the preset judgment threshold are regarded as impacted wisdom teeth in the patient's oral cavity 3D model.

[0032] Furthermore, the acquisition of the analysis region in each oral CT image includes:

[0033] Edge detection is performed on each oral CT image to obtain edge pixels. Curve fitting is then performed on the edge pixels to obtain the edge lines in each oral CT image. The closed region formed by each edge line is taken as the analysis region in each oral CT image.

[0034] Furthermore, obtaining the reference teeth for each wisdom tooth in the patient's three-dimensional oral model includes:

[0035] The FDI tooth position recording method was used to mark the teeth in the patient's three-dimensional oral model, and the teeth in the same quadrant as each wisdom tooth were recorded as the control teeth for each wisdom tooth.

[0036] Furthermore, the edge detection for each oral CT image is performed using the Sobel operator.

[0037] The present invention has the following beneficial effects:

[0038] Firstly, artifacts in CT images can interfere with the structure of tooth tissue and increase the complexity of the image. To improve the accuracy of subsequent analysis, the shape features, edge clarity features, and gray-scale distribution features of the tooth and artifact parts are combined to filter the oral CT images, removing the artifact parts and obtaining enhanced images of the oral CT images. Analyzing the enhanced images can effectively reduce the interference of artifacts on the process of identifying impacted wisdom teeth and improve the accuracy of impacted wisdom tooth identification.

[0039] Secondly, impacted wisdom teeth differ significantly from normal wisdom teeth in terms of shape regularity, size, and growth direction. These characteristics are analyzed using sagittal, axial, and coronal images of the oral cavity to determine the likelihood of an impacted wisdom tooth in the 3D model, thus confirming the patient's impacted wisdom tooth. Combining the analysis of features of impacted wisdom teeth in enhanced images from different perspectives of oral CT scans effectively addresses the issue of low accuracy in identifying impacted wisdom teeth due to errors in data from a single perspective, further improving the accuracy of impacted wisdom tooth identification. Attached Figure Description

[0040] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 The flowchart illustrates the steps of an impacted tooth identification method based on three-dimensional oral imaging, as provided in one embodiment of the present invention.

[0042] Figure 2 A flowchart illustrating the steps of an enhanced image acquisition method for oral CT images provided in an embodiment of the present invention;

[0043] Figure 3 This is a flowchart illustrating the steps of a method for determining impacted wisdom teeth according to an embodiment of the present invention.

[0044] Figure 4 This is a system structure diagram of an impacted tooth assisted recognition system based on three-dimensional oral imaging, provided in one embodiment of the present invention.

[0045] Figure 5 This is a schematic diagram of a computer device for assisting in the identification of impacted teeth based on three-dimensional oral imaging, as provided in one embodiment of the present invention. Detailed Implementation

[0046] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method for assisting in the identification of impacted teeth based on three-dimensional oral imaging proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0048] The following description, in conjunction with the accompanying drawings, details the specific scheme of the impacted tooth assisted identification method based on three-dimensional oral imaging provided by the present invention.

[0049] Example 1:

[0050] This invention proposes a method for assisted identification of impacted teeth based on three-dimensional oral imaging. Please refer to [link to relevant documentation]. Figure 1 The diagram illustrates a flowchart of a method for assisting in the identification of impacted teeth based on three-dimensional oral imaging, according to an embodiment of the present invention. The method includes:

[0051] Step S1: Acquire the patient's oral CT images, including: axial CT images, coronal CT images, and sagittal CT images.

[0052] Specifically, a cone-beam computed tomography (CBCT) system is used to scan the patient's oral cavity, obtaining oral computed tomography (CT) images. Oral CT images include axial CT images, coronal CT images, and sagittal CT images. It should be noted that CT images are typically grayscale images.

[0053] Step S2: Obtain the analysis region in each oral CT image; based on the grayscale change of each edge pixel in each oral CT image, the shape, edge clarity and grayscale distribution of the analysis region where each edge pixel is located, filter the edge pixels in each oral CT image to obtain the enhanced image of each oral CT image.

[0054] In oral CT images, artifacts often interfere with real structures and may even obscure actual lesions or structures, affecting the doctor's interpretation of the CT images. Furthermore, artifacts increase image complexity during feature extraction, impacting subsequent analysis and processing. Therefore, artifact removal processing is necessary for oral CT images.

[0055] First, the analysis region in the oral CT image is obtained. The analysis region may be the tooth area and the artifact area. In this embodiment, the method for obtaining the analysis region is as follows: edge detection is performed on each oral CT image to obtain edge pixels, and curve fitting is performed on the edge pixels to obtain the edge lines in each oral CT image; the closed region formed by each edge line is taken as the analysis region in each oral CT image.

[0056] It should be noted that the Sobel operator was selected for edge detection of the oral CT images, and the least squares method was used for curve fitting of the edge pixels in the oral CT images. Both the Sobel operator and the least squares method are techniques well-known to those skilled in the art and will not be elaborated upon here.

[0057] The characteristics of teeth in CT images are: teeth exhibit clearly regular structural features, relatively uniform grayscale distribution, and a clear boundary with surrounding tissues. The characteristics of artifacts in CT images are: artifacts may appear as irregular shadows or stripes, with uneven grayscale distribution and relatively blurred boundaries with surrounding tissues. The edge sharpness of the analysis area reflects the clarity of the boundary between the analysis area and surrounding tissues, and the grayscale variation of edge pixels reflects the sharpness of the local edge position of the analysis area. Therefore, the grayscale variation of edge pixels, the shape of the analysis area where the edge pixels are located, the edge sharpness, and the grayscale distribution can clearly distinguish the characteristics of teeth and artifacts, determine the probability that the edge pixels are located in the tooth area, and thus filter the oral CT image to remove artifacts, obtaining an enhanced image of the oral CT image.

[0058] Please see Figure 2 The diagram illustrates a flowchart of a method for acquiring enhanced images of an oral CT image according to an embodiment of the present invention. The method includes:

[0059] Step S210: Based on the grayscale change of each edge pixel in each oral CT image, the shape of the analysis area where each edge pixel is located, the edge clarity and grayscale distribution, obtain the possible tooth values ​​of each edge pixel in each oral CT image.

[0060] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the possible values ​​of teeth includes: for each oral CT image, obtaining the shape grayscale index of each analysis region based on the shape and grayscale distribution of each analysis region in the oral CT image; calculating the mean of the absolute values ​​of the differences between the gradient values ​​of two adjacent edge pixels on the edge of each analysis region in the oral CT image as the edge clarity index of each analysis region; and obtaining the possible values ​​of teeth for each edge pixel in the oral CT image based on the gradient value of each edge pixel in the oral CT image, as well as the shape grayscale index and edge clarity index of the analysis region where each edge pixel is located.

[0061] In this embodiment, the method for obtaining the shape grayscale index includes: for each analysis region in each oral CT image, obtaining the discrete index of the grayscale values ​​of all pixels in the analysis region; calculating the concentrated value of the grayscale values ​​of all pixels in the analysis region; obtaining the geometric centroid and grayscale centroid of the analysis region, and using the distance between the geometric centroid and the grayscale centroid as the shape rule value; obtaining the shape grayscale index of the analysis region based on the discrete index, the concentrated value, and the shape rule value; the discrete index and the shape rule value are both negatively correlated with the shape grayscale index, and the concentrated value is positively correlated with the shape grayscale index.

[0062] The dispersion index of grayscale values ​​of all pixels within the analysis area reflects the degree of grayscale dispersion within the analysis area and is used to measure the uniformity of grayscale distribution within the analysis area. The variance, standard deviation, interquartile range, and range of a set of data can all reflect the degree of dispersion of the data. In this embodiment, variance is selected as the dispersion index, that is, the variance of grayscale values ​​of pixels within the analysis area is used as the dispersion index. The smaller the dispersion index, the more uniform the grayscale distribution within the analysis area. The larger the shape grayscale index, the greater the probability that the analysis area represents a tooth.

[0063] Given that the teeth are relatively clear while the artifacts are relatively blurry, the central tendency (CLT) of the grayscale values ​​of the pixels within the analysis area is used to measure the clarity of that area. The CLT represents the overall grayscale level of the pixels within the analysis area. The mean, median, and mode can all represent the overall level of a set of data. In this embodiment, the mean is chosen as the CLT, meaning the average grayscale value of the pixels within the analysis area is used as the CLT. A larger CLT indicates a clearer analysis area, a higher shape grayscale index, and a greater likelihood that the analysis area represents the teeth.

[0064] The gray-level centroid of the analysis region represents the center of the gray-level distribution within that region. The distance between the geometric centroid and the gray-level centroid of the analysis region, known as the shape regularity value, measures the degree of shape regularity of the analysis region. A smaller shape regularity value indicates more uniform gray-level distribution within the analysis region, meaning the gray-level distribution better matches the geometric shape, and the greater the likelihood of symmetry. In other words, the more regular the shape of the analysis region, the higher the shape gray-level index, and the greater the likelihood that the analysis region represents a tooth. Conversely, a larger shape regularity value indicates significant brightness differences within the analysis region and greater complexity in its shape, meaning the more irregular the shape of the analysis region, and the greater the likelihood that the analysis region represents an artifact. The methods for obtaining the geometric centroid and gray-level centroid of the region are well-known techniques and will not be elaborated upon here.

[0065] In summary, both the discrete index and the shape regularity value are negatively correlated with the shape grayscale index, while the lumped value is positively correlated with the shape grayscale index. In this embodiment of the invention, the product of the discrete index and the shape regularity value of the analysis region is used as the numerator, and the lumped value is used as the denominator to obtain the ratio, which is then negatively correlated to obtain the shape regularity value of the analysis region. If the shape regularity value is larger, the analysis region is more likely to represent a tooth region, and thus the tooth potential value is larger.

[0066] In the embodiments of the present invention, the correlation between discrete indices, shape regularity values ​​and concentrated values ​​and shape grayscale indices can also be constructed through other basic mathematical operations, which are not limited or elaborated here.

[0067] Because tooth tissue has a clear boundary with surrounding tissue, while the boundary between the artifact region and surrounding tissue is relatively blurred, the gradient values ​​of edge pixels in the tooth region are generally larger, while the gradient values ​​of edge pixels in the artifact region are smaller and more random. Therefore, the gradient value difference between adjacent edge pixels on the tooth region is smaller, while the gradient value difference between adjacent edge pixels on the artifact region is larger. Thus, the mean of the absolute values ​​of the differences between the gradient values ​​of two adjacent edge pixels on the edge of the analysis region can measure the edge sharpness of the analysis region, yielding an edge sharpness index. The smaller the edge sharpness index, the clearer the boundary between the analysis region and surrounding tissue, the greater the likelihood that the analysis region represents a tooth region, and the higher the tooth probability value.

[0068] The gradient value of an edge pixel represents the sharpness of the local edge. The larger the gradient value of an edge pixel, the sharper the local edge, and the greater the probability that the edge pixel is located at the edge of the tooth region, thus the greater the tooth probability value.

[0069] Therefore, the gradient value of edge pixels and the shape grayscale index are both positively correlated with the possible value of teeth, while the edge clarity index is negatively correlated with the possible value of teeth.

[0070] In one specific implementation of this invention, the possible values ​​of teeth are expressed by the formula:

[0071] p a =Norm(G a ×exp(-BQ a )×XH a )

[0072]

[0073] In the formula, p a G represents the possible values ​​of the tooth at the a-th edge pixel in each oral CT image; a BQ represents the gradient value of the a-th edge pixel in each oral CT image. aXH is the edge clarity index of the analysis region where the a-th edge pixel is located in each oral CT image; a σ represents the shape grayscale index of the analysis region where the a-th edge pixel is located in each oral CT image; a It is a discrete index of the grayscale value of the pixel within the analysis region where the a-th edge pixel is located in each oral CT image; D represents the concentrated grayscale values ​​of pixels within the analysis region where the a-th edge pixel is located in each oral CT image; a Let be the shape regularity value of the analysis region where the a-th edge pixel is located in each oral CT image; Norm is the normalization function; exp is the exponential function with the natural constant as the base. It should be noted that if the tooth may have values ​​p... a The larger the value, the greater the probability that the a-th edge pixel is located at the edge of the tooth tissue.

[0074] Step S220: For each oral CT image, select one edge pixel in the oral CT image as the point to be filtered, use the possible values ​​of the teeth as the weights of the gray values ​​of the remaining edge pixels in the preset filtering window of the point to be filtered, and obtain the weighted average value as the filtered gray value of the point to be filtered.

[0075] Each edge pixel is filtered by using other edge pixels within its local region. The tooth values ​​of the edge pixels within the local region of the tooth part may be large, resulting in a large grayscale value after filtering. However, the tooth values ​​of the edge pixels within the local region of the artifact part may be small, resulting in a smaller grayscale value after filtering. This achieves the purpose of removing artifacts from oral CT images.

[0076] Taking any edge pixel in an oral CT image as the point to be filtered as an example, the filtered gray value of the point to be filtered is expressed by the formula:

[0077]

[0078] In the formula, I g ρ is the grayscale value of the point to be filtered; g is the point to be filtered; W is the total number of edge pixels other than the point to be filtered within the preset filtering window; ρ g,w I represents the possible tooth values ​​of the w-th edge pixel (excluding the pixel itself) within the preset filtering window of the point to be filtered; w This is the grayscale value of the w-th edge pixel (excluding the point to be filtered) within the preset filtering window of the point to be filtered.

[0079] It should be noted that in this embodiment, the edge pixels are located at the center of their preset filtering window. The size of the preset filtering window is taken as an empirical value of 9×9, and the implementer can set it according to the specific situation. The method for obtaining the filtered grayscale values ​​of all edge pixels in each oral CT image is the same as the method for obtaining the filtered grayscale values ​​of the points to be filtered.

[0080] Step S230: Update the gray values ​​of edge pixels in the oral CT image using filtered gray values ​​to obtain an enhanced image of the oral CT image.

[0081] In each oral CT image, the gray values ​​of non-edge pixels remain unchanged, while the gray values ​​of edge pixels are replaced with filtered gray values ​​to obtain an enhanced image of the oral CT image. The enhanced image effectively removes the influence of artifacts and preserves the features of the teeth.

[0082] Step S3: Obtain the tooth region in the enhanced image of each oral CT image; perform three-dimensional reconstruction on the enhanced images of all oral CT images to obtain a three-dimensional oral model, and determine the wisdom teeth in the three-dimensional oral model; based on the shape difference and tilt difference between the corresponding tooth region of each wisdom tooth in the enhanced image of the sagittal CT image of the oral model and the other tooth regions, the area of ​​the corresponding tooth region of each wisdom tooth in the enhanced image of the axial CT image of the oral model, and the distance between adjacent tooth regions in the enhanced image of the coronal CT image of the oral model, determine the impacted wisdom teeth in the patient's three-dimensional oral model.

[0083] To analyze tooth features, the tooth region is obtained from the enhanced image of the oral CT image. It should be noted that in this embodiment, the method for obtaining the tooth region is the same as the method for obtaining the analysis region. However, since the enhanced image of the oral CT image has effectively removed the influence of artifacts, the region obtained by edge detection in the enhanced image is only the tooth region. Other embodiments may also use methods such as contour extraction and region growing to obtain the tooth region.

[0084] A three-dimensional oral model of the patient is obtained by CT reconstruction using enhanced images of the patient's axial, coronal, and sagittal CT images. The teeth in the three-dimensional oral model are marked using the FDI (World Dental Federation) tooth position recording system. In this embodiment, teeth numbered 18, 28, 38, and 48 are designated as wisdom teeth. This method only identifies impacted wisdom teeth. The FDI tooth position recording system is well-known and will not be described in detail here.

[0085] Impacted wisdom teeth may present as irregular shapes, be small in size, and grow at an angle, while normal wisdom teeth are more regular in shape, larger in size, and grow vertically. In addition, due to the pressure from surrounding teeth or bones, the distance between impacted wisdom teeth and their adjacent teeth is smaller than the distance between normal adjacent teeth. Therefore, we analyze the shape differences and degree of tilt between the corresponding tooth region and the other tooth regions in the enhanced images of sagittal CT images of each wisdom tooth, and combine the area of ​​the corresponding tooth region in the enhanced images of axial CT images of the wisdom tooth with the distance between adjacent tooth regions in the enhanced images of coronal CT images of the wisdom tooth to determine the impacted wisdom teeth.

[0086] Please see Figure 3 The diagram illustrates a flowchart of a method for determining impacted wisdom teeth according to an embodiment of the present invention, the method comprising:

[0087] Step S310: Obtain the control tooth for each wisdom tooth in the patient's three-dimensional oral model; based on the shape difference and tilt difference of the corresponding tooth region in the enhanced image of the sagittal CT image of the oral cavity, obtain the growth deviation value of each wisdom tooth.

[0088] The FDI tooth position recording method divides the teeth in the oral cavity into four quadrants, and the teeth in the same quadrant as each wisdom tooth are designated as the control teeth for that wisdom tooth. As an example, in the upper right quadrant, teeth numbered 11, 12, 13, 14, 15, 16, and 17 are the control teeth for tooth number 18, and each wisdom tooth has 7 control teeth.

[0089] Impacted wisdom teeth are restricted in their growth due to pressure from surrounding teeth or bone, resulting in irregular shapes, small sizes, and even tilted growth. Normal wisdom teeth, on the other hand, have ample space for growth, leading to more regular shapes, larger sizes, and vertical growth. This study assumes that the control teeth for each wisdom tooth are in good condition, meaning they are relatively regular in size and grow perpendicular to the horizontal plane. The sagittal image provides a lateral view of the oral structure, clearly showing the vertical growth of the teeth.

[0090] The shape and tilt of the corresponding tooth region in the enhanced image of the oral sagittal CT image of the control tooth represent the normal tooth growth status. The difference in shape and tilt of the corresponding tooth region between each wisdom tooth and its control tooth in the enhanced image of the oral sagittal CT image in the three-dimensional oral model represents the degree of deviation of the wisdom tooth from the normal tooth growth status, and the growth deviation value of the wisdom tooth is obtained.

[0091] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the growth deviation value of each wisdom tooth includes: obtaining the minimum bounding rectangle of each tooth region in the enhanced image of the oral sagittal CT image; recording the angle between the direction of the long side of the minimum bounding rectangle of each tooth region and a preset direction as the direction index of the corresponding tooth region; calculating the absolute value of the difference between the length and width of the minimum bounding rectangle of each tooth region as the shape index of each tooth region; obtaining the local normal deviation value of each wisdom tooth and each control tooth in the enhanced image of the oral sagittal CT image based on the differences in the direction index and shape index of the corresponding tooth region in the three-dimensional model of the patient's oral cavity; and summing the local growth deviation values ​​of each wisdom tooth and all control teeth in the three-dimensional model of the oral cavity as the growth deviation value of each wisdom tooth.

[0092] In sagittal images of the oral cavity, the direction of the long side of the minimum bounding rectangle of each tooth region is usually close to the vertical direction. The directional index of the tooth region is used to measure the growth direction of the tooth, i.e., the degree of inclination. Since the shape of normal wisdom teeth is relatively regular, the length and width of the minimum bounding rectangle of the tooth region are relatively close. Impacted wisdom teeth may have irregular shapes, so the length and width of the minimum bounding rectangle of impacted wisdom teeth differ more. Therefore, in this embodiment, the absolute value of the difference between the length and width of the minimum bounding rectangle of each tooth region in the enhanced image of the oral sagittal CT image is used as a shape index to measure the regularity of the shape of the tooth region.

[0093] The greater the differences in both directional and shape indices between each wisdom tooth and its corresponding tooth region in the enhanced sagittal CT image of the oral cavity, the greater the difference in shape and growth direction compared to normal teeth, the larger the local deviation value, and the greater the likelihood that the wisdom tooth is impacted. Therefore, in this embodiment of the invention, the product of the differences in directional and shape indices between each wisdom tooth and its corresponding tooth region in the enhanced sagittal CT image of the oral cavity is used as the local deviation value. The deviation between the wisdom tooth and all its control teeth is analyzed to obtain the growth deviation value.

[0094] It should be noted that in this embodiment of the invention, the preset direction is horizontal, and the value range of the direction index of the tooth region is [0°, 90°].

[0095] In one specific implementation of this invention, the growth deviation value of each wisdom tooth is expressed by the formula:

[0096]

[0097] In the formula, SP represents the growth deviation value of each wisdom tooth in the patient's three-dimensional oral model; θ0 represents the orientation index of the corresponding tooth region in the enhanced image of the sagittal CT image of the patient's three-dimensional oral model; θ m Xm represents the orientation index of the corresponding tooth region in the enhanced sagittal CT image of the m-th control tooth of each wisdom tooth in the patient's 3D oral model; X0 represents the shape index of the corresponding tooth region in the enhanced sagittal CT image of each wisdom tooth in the patient's 3D oral model; Xm ... m The shape index of the corresponding tooth region in the enhanced image of the sagittal CT image of the m-th control tooth of each wisdom tooth in the patient's three-dimensional oral model; |θ0-θ m |×|X0-X m | represents the local normal deviation between each wisdom tooth and its m-th control tooth in the patient's 3D oral model; M represents the total number of control teeth for each wisdom tooth in the patient's 3D oral model; || is the absolute value function; Norm is the normalization function.

[0098] Step S320: Obtain the distance between two adjacent tooth regions in the enhanced image of the coronal CT image of the oral cavity; construct the tooth distance set for each wisdom tooth by the distance between the two adjacent tooth regions in the enhanced image of the coronal CT image of each control tooth in the oral cavity 3D model; and sum the absolute values ​​of the differences between the distance between the two adjacent tooth regions in the enhanced image of the coronal CT image of each wisdom tooth in the oral cavity 3D model and each element in the tooth distance set as the normal distance difference value for each wisdom tooth.

[0099] MydentalX software was used to measure the distance between two adjacent tooth regions in enhanced images of coronal CT images of the oral cavity.

[0100] Taking the upper right quadrant of the oral cavity as an example for analysis, teeth numbered 11, 12, 13, 14, 15, 16 and 17 are the control teeth for tooth number 18. The numbers of adjacent teeth in all the control teeth of wisdom tooth number 18 are: (11, 12), (12, 13), (13, 14), (14, 15), (15, 16), (16, 17). The spacing between the above two adjacent teeth constitutes the tooth spacing set of wisdom tooth number 18.

[0101] The coronal view provides an anterior-posterior cross-sectional view of the oral cavity; because impacted wisdom teeth are compressed by surrounding teeth or bones, the distance between impacted wisdom teeth and their adjacent teeth is smaller than the distance between normal adjacent teeth. The normal distance difference value is determined based on the two types of distance differences, and the probability of wisdom teeth being impacted is measured by the distance between teeth.

[0102] The distance between any two adjacent teeth in all control teeth for each wisdom tooth is considered the normal tooth spacing. The greater the difference between the distance between each wisdom tooth and its adjacent teeth and the elements in the tooth spacing set, the greater the likelihood that each wisdom tooth is impacted. Therefore, the sum of the absolute values ​​of the differences between the distances between the corresponding two tooth regions in the enhanced coronal CT image of each wisdom tooth and its adjacent teeth in the 3D oral model and each element in the tooth spacing set is used as the normal spacing difference value for each wisdom tooth. A larger normal spacing difference value indicates a greater difference between the distance between each wisdom tooth and its adjacent teeth and the normal tooth spacing, thus increasing the likelihood that each wisdom tooth is impacted.

[0103] It should be noted that in the 3D model of the oral cavity, each wisdom tooth has only one adjacent tooth. The adjacent tooth of wisdom tooth number 18 is numbered 17, the adjacent tooth of wisdom tooth number 28 is numbered 27, the adjacent tooth of wisdom tooth number 38 is numbered 37, and the adjacent tooth of wisdom tooth number 48 is numbered 47.

[0104] Step S330: Based on the area of ​​the corresponding tooth region in the enhanced image of the axial CT image of each wisdom tooth in the oral cavity 3D model, the growth deviation value and the difference value between the normal interval, obtain the impaction characteristic value of each wisdom tooth; the area and the impaction characteristic value are negatively correlated, while the growth deviation value and the difference value between the normal interval are positively correlated with the impaction characteristic value; use the impaction characteristic value to determine the impacted wisdom teeth among all wisdom teeth in the patient's oral cavity 3D model.

[0105] Impacted wisdom teeth are partially or completely buried under the gum line due to insufficient space, incorrect angle, or interference from other teeth. Therefore, the occlusal surface of impacted wisdom teeth is usually smaller. Normal wisdom teeth usually erupt completely, so the occlusal surface of impacted wisdom teeth is usually smaller than that of normal wisdom teeth. Axial images provide a horizontal view of the teeth, clearly showing the occlusal surface. In contrast, the area corresponding to the wisdom tooth in the enhanced image of an axial CT scan of the oral cavity represents the occlusal surface of the wisdom tooth. The smaller the area of ​​this occlusal surface, the more likely the wisdom tooth is not fully erupted and is therefore impacted.

[0106] Therefore, the area of ​​the corresponding tooth region in the enhanced image of an axial CT scan of the oral cavity is negatively correlated with the impaction characteristic value, while the growth deviation value and the difference value between the normal interval and the impaction characteristic value are positively correlated. The larger the impaction characteristic value of the wisdom tooth, the greater the likelihood that the wisdom tooth is impacted.

[0107] In this embodiment of the invention, the area of ​​the corresponding tooth region in the enhanced image of the axial CT image of the oral cavity of each wisdom tooth in the three-dimensional model of the patient's oral cavity is negatively correlated and mapped. The product of the growth deviation value, the difference value of the normal interval and the negative correlation mapping result is normalized to obtain the impaction characteristic value of each wisdom tooth.

[0108] In this embodiment of the invention, other basic mathematical operations can also be used to construct the correlation between the area of ​​the corresponding tooth region, the growth deviation value and the difference value of the normal interval in the enhanced image of the axial CT image of each wisdom tooth, which is not limited or elaborated here.

[0109] It should be noted that in this embodiment of the invention, the Norm function is used for normalization, and the inverse of the area of ​​the tooth region is taken to achieve negative correlation mapping; alternatively, function transformation, min-max normalization, or other normalization methods can be used to take negative numbers and perform negative correlation mapping through function transformation, which is not limited here. In this embodiment, the total number of pixels in the tooth region is taken as the area of ​​the tooth region.

[0110] For all wisdom teeth in the patient's 3D oral model, the wisdom teeth with impacted characteristic values ​​exceeding a preset threshold are considered impacted wisdom teeth in the patient's 3D oral model. It should be noted that in this embodiment, the preset threshold is an empirical value of 0.7, and implementers can set it according to specific circumstances.

[0111] To facilitate the identification of impacted wisdom teeth in other patients, the patient's oral cavity 3D model is input into a pre-trained neural network, which outputs the number of the impacted wisdom teeth in the oral cavity 3D model.

[0112] This invention uses a convolutional neural network to identify impacted wisdom teeth in a three-dimensional oral cavity model. The input to the neural network is the patient's three-dimensional oral cavity model, and the output is the number of the impacted wisdom teeth in the three-dimensional oral cavity model.

[0113] The relevant content regarding convolutional neural networks includes: the neural network dataset is divided into a training set and a validation set; the training process of the neural network is the identification process of impacted wisdom teeth in a three-dimensional oral cavity model, specifically: obtaining the number of the impacted wisdom teeth in the three-dimensional oral cavity model; the loss function of the neural network is the cross-entropy function. Convolutional neural networks are well-known technology to those skilled in the art and will not be elaborated upon here. It should be noted that the number of the impacted wisdom teeth refers to the tooth number in the FDI tooth position recording method.

[0114] This invention is now complete.

[0115] Example 2:

[0116] This invention proposes an impacted tooth recognition system based on three-dimensional oral imaging. Please refer to [link / reference]. Figure 4 The diagram illustrates a system structure of an impacted tooth assisted recognition system based on three-dimensional oral imaging, according to an embodiment of the present invention. The system includes:

[0117] Data acquisition module 410 is used to acquire oral CT images of a patient, the oral CT images including: oral axial CT images, oral coronal CT images and oral sagittal CT images;

[0118] The image enhancement module 420 is used to acquire the analysis region in each oral CT image; based on the gray level change of each edge pixel in each oral CT image, the shape, edge clarity and gray level distribution of the analysis region where each edge pixel is located, the edge pixels in each oral CT image are filtered to obtain the enhanced image of each oral CT image.

[0119] The impacted wisdom tooth recognition module 430 is used to acquire the tooth region in the enhanced image of each oral CT image; to perform three-dimensional reconstruction on the enhanced images of all oral CT images to obtain a three-dimensional oral model, and to identify the wisdom teeth in the three-dimensional oral model; based on the shape difference and tilt difference between the corresponding tooth region of each wisdom tooth in the enhanced image of the sagittal CT image of the oral model and the other tooth regions, the area of ​​the corresponding tooth region of each wisdom tooth in the enhanced image of the axial CT image of the oral model, and the distance between adjacent tooth regions in the enhanced image of the coronal CT image of the oral model, the impacted wisdom teeth in the patient's three-dimensional oral model are determined.

[0120] It should be noted that the devices provided in the above embodiments are only illustrative examples of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the impacted tooth assisted recognition system based on three-dimensional oral imaging and the impacted tooth assisted recognition method based on three-dimensional oral imaging provided in the above embodiments belong to the same concept, and their specific implementation process is detailed in the method embodiments, which will not be repeated here.

[0121] Example 3:

[0122] Figure 5 This is a schematic diagram of a computer device for assistive recognition of impacted teeth based on three-dimensional oral imaging, as provided in one embodiment of the present invention. For example,... Figure 5 As shown, the computer device includes: a memory 501, a processor 502, and a computer program 503 stored in the memory 501 and running on the processor 502, wherein when the processor 502 executes the computer program 503, the computer device can execute any of the aforementioned methods for assisting in the identification of impacted teeth based on three-dimensional images of the oral cavity.

[0123] Furthermore, embodiments of this application also protect an apparatus that may include a memory and a processor, wherein the memory stores executable program code, and the processor is used to call and execute the executable program code to perform an impacted tooth assisted identification method based on three-dimensional oral imaging provided in embodiments of this application.

[0124] This embodiment can divide the device into functional modules based on the above method example. For example, each module can correspond to a separate function, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0125] It should be understood that the device provided in this embodiment is used to perform the above-described method for assisting in the identification of impacted teeth based on three-dimensional oral imaging, and therefore can achieve the same effect as the above-described implementation method.

[0126] When using integrated units, the device may include a processing module and a storage module. When applied to a workpiece, the processing module can be used to control and manage the workpiece's operations. The storage module can be used to support the execution of program code by the workpiece.

[0127] The processing module may be a processor or a controller, which can implement or execute various exemplary logic blocks, modules, and circuits contained in conjunction with the disclosure of this application. The processor may also be a combination of functions that implement computing capabilities, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and microprocessors, etc., and the storage module may be a memory.

[0128] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0129] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0130] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for auxiliary recognition of impacted teeth based on three-dimensional images of the oral cavity, characterized in that, The method includes: Acquire oral CT images of the patient, including: axial CT images, coronal CT images, and sagittal CT images; The analysis region in each oral CT image is obtained; based on the gray level change of each edge pixel in each oral CT image, the shape, edge clarity and gray level distribution of the analysis region where each edge pixel is located, the edge pixels in each oral CT image are filtered to obtain the enhanced image of each oral CT image. Obtain the tooth region in the enhanced image of each oral CT image; perform 3D reconstruction on the enhanced images of all oral CT images to obtain a 3D oral model, and identify the wisdom teeth in the 3D oral model; based on the shape and tilt differences between the corresponding tooth region of each wisdom tooth in the enhanced image of the sagittal CT image of the oral model and the other tooth regions, the area of ​​the corresponding tooth region of each wisdom tooth in the enhanced image of the axial CT image of the oral model, and the distance between adjacent tooth regions in the enhanced image of the coronal CT image of the oral model, determine the impacted wisdom teeth in the patient's 3D oral model; The method for acquiring the enhanced image of each oral CT image includes: Based on the degree of grayscale change of each edge pixel in each oral CT image, the shape, edge clarity and grayscale distribution of the analysis area where each edge pixel is located, the possible tooth values ​​of each edge pixel in each oral CT image are obtained. For each oral CT image, one edge pixel in the oral CT image is randomly selected and recorded as the point to be filtered. The possible values ​​of the tooth are used as the weights of the gray values ​​of the remaining edge pixels excluding the point to be filtered within the preset filtering window of the point to be filtered. The weighted average value is then used as the filtered gray value of the point to be filtered. The gray values ​​of edge pixels in the oral CT image are updated using the filtered gray values ​​to obtain an enhanced image of the oral CT image. The process of obtaining the possible tooth values ​​for each edge pixel in each oral CT image includes: For each oral CT image, the shape grayscale index of each analysis region is obtained based on the shape and grayscale distribution of each analysis region in the oral CT image; The mean of the absolute values ​​of the differences in gradient values ​​between two adjacent edge pixels on the edge of each analysis region in an oral CT image is calculated as the edge clarity index for each analysis region. Based on the gradient value of each edge pixel in the oral CT image, as well as the shape grayscale index and the edge clarity index of the analysis region where each edge pixel is located, the possible tooth values ​​of each edge pixel in the oral CT image are obtained.

2. The method for assisted identification of impacted teeth based on three-dimensional oral imaging according to claim 1, characterized in that, The process of obtaining the shape grayscale index of each analysis region includes: For each analysis region in each oral CT image, obtain the discrete index of the gray values ​​of all pixels in the analysis region; calculate the central value of the gray values ​​of all pixels in the analysis region. Obtain the geometric centroid and grayscale centroid of the analysis region, and use the distance between the geometric centroid and the grayscale centroid as the shape rule value; Based on the discrete index, the setpoint, and the shape rule value, the shape grayscale index of the analysis area is obtained; the discrete index and the shape rule value are both negatively correlated with the shape grayscale index, and the setpoint is positively correlated with the shape grayscale index.

3. The method for assisted identification of impacted teeth based on three-dimensional oral imaging according to claim 1, characterized in that, The method of determining impacted wisdom teeth in the patient's oral 3D model based on the shape and tilt differences between the corresponding tooth region and other tooth regions in the enhanced sagittal CT image of each wisdom tooth in the oral 3D model, the area of ​​the corresponding tooth region in the enhanced axial CT image of each wisdom tooth in the oral 3D model, and the distance between adjacent tooth regions in the enhanced coronal CT image of the oral 3D model includes: Obtain the control tooth for each wisdom tooth in the patient's three-dimensional oral model; based on the shape difference and tilt difference of the corresponding tooth region in the enhanced image of the sagittal CT image of the oral cavity, obtain the growth deviation value of each wisdom tooth. The distance between two adjacent tooth regions in the enhanced image of the coronal CT image of the oral cavity is obtained; the tooth distance set of each wisdom tooth is constructed by the distance between the two adjacent tooth regions in the enhanced image of the coronal CT image of each control tooth in the oral cavity 3D model; the sum of the absolute values ​​of the differences between the distance between the two adjacent tooth regions in the enhanced image of the coronal CT image of each wisdom tooth in the oral cavity 3D model and each element in the tooth distance set is used as the normal distance difference value of each wisdom tooth. Based on the area of ​​the corresponding tooth region in the enhanced image of the axial CT image of each wisdom tooth in the three-dimensional oral model, the growth deviation value and the difference value between the normal spacing are used to obtain the impaction characteristic value of each wisdom tooth; the area and the impaction characteristic value are negatively correlated, and the growth deviation value and the difference value between the normal spacing are both positively correlated with the impaction characteristic value. The impacted wisdom teeth in the patient's three-dimensional oral cavity model were determined using the impacted characteristic values.

4. The method for assisted identification of impacted teeth based on three-dimensional oral imaging according to claim 3, characterized in that, The process of obtaining the growth deviation value of each wisdom tooth includes: Obtain the minimum bounding rectangle of each tooth region in the enhanced image of the sagittal CT image of the oral cavity; The angle between the direction of the long side of the minimum bounding rectangle of each tooth region and the preset direction is recorded as the direction index of the corresponding tooth region; the absolute value of the difference between the length and width of the minimum bounding rectangle of each tooth region is calculated as the shape index of each tooth region. Based on the differences in the directional and shape indices of the corresponding tooth regions in the enhanced sagittal CT images of the patient's oral cavity 3D model, the local normal deviation value of each wisdom tooth and each control tooth is obtained; the sum of the local normal deviation values ​​of each wisdom tooth and all control teeth in the oral cavity 3D model is taken as the growth deviation value of each wisdom tooth.

5. The method for assisted identification of impacted teeth based on three-dimensional oral imaging according to claim 3, characterized in that, The process of using the impacted characteristic value to determine the impacted wisdom teeth among all wisdom teeth in the patient's three-dimensional oral model includes: For all wisdom teeth in the patient's oral cavity 3D model, the wisdom teeth with impacted characteristic values ​​greater than the preset judgment threshold are regarded as impacted wisdom teeth in the patient's oral cavity 3D model.

6. The method for assisted identification of impacted teeth based on three-dimensional oral imaging according to claim 1, characterized in that, The process of acquiring the analysis region in each oral CT image includes: Edge detection is performed on each oral CT image to obtain edge pixels. Curve fitting is then performed on the edge pixels to obtain the edge lines in each oral CT image. The closed region formed by each edge line is taken as the analysis region in each oral CT image.

7. The method for assisted identification of impacted teeth based on three-dimensional oral imaging according to claim 3, characterized in that, The process of obtaining the reference teeth for each wisdom tooth in the patient's three-dimensional oral model includes: The FDI tooth position recording method was used to mark the teeth in the patient's three-dimensional oral model, and the teeth in the same quadrant as each wisdom tooth were recorded as the control teeth for each wisdom tooth.

8. The method for assisted identification of impacted teeth based on three-dimensional oral imaging according to claim 6, characterized in that, The edge detection performed on each oral CT image is performed using the Sobel operator.

Images