CBCT feature extraction method for evaluating bone density of alveolar bone

Abstract

The invention relates to the technical field of medical auxiliary diagnosis, in particular to a CBCT (Cone Beam Computed Tomography) feature extraction method for evaluating bone density of alveolar bone. According to the method, the sub-blocks in the alveolar bone block are combined based on the difference of the absorption degradation significance of different sub-blocks, different connection blocks are generated, and the lesion block is selected; according to the boundary clearness degree of all lesion blocks in the alveolar bone block and the lesion degree difference condition of different sub-blocks in the alveolar bone block, obtaining the lesion activity degree; according to the absorption degradation significance and fluctuation degree of the sub-blocks of the alveolar bone block in different preset directions, acquiring bone density spatial dispersion; and according to the lesion activity degree and the bone density spatial dispersion degree of all the alveolar bone blocks in the scanning time window and the correlation degree of the lesion activity degree and the bone density spatial dispersion degree, acquiring the bone pathology risk degree. According to the method, the analysis level of the CBCT image is improved from morphological observation to pathological function inference, and the reliability of alveolar bone bone density evaluation is improved.

Description

Technical Field

[0001] This invention relates to the field of medical auxiliary diagnostic technology, specifically to a CBCT feature extraction method for alveolar bone mineral density assessment. Background Technology

[0002] In dental clinical practice, accurate assessment of alveolar bone mineral density is a crucial basis for developing treatment plans for dental implants, orthodontic treatment, and periodontal disease. Cone-beam computed tomography (CBCT) is characterized by its flexible scanning range, low radiation dose, and high spatial resolution, providing three-dimensional alveolar bone structural information and has become an important imaging tool for alveolar bone assessment.

[0003] Existing CBCT-based bone mineral density assessment methods are mostly limited to basic morphological parameters such as grayscale value and volume. However, the correlation between single grayscale value or morphological parameters and pathology is weak, and their extraction is easily affected by CBCT imaging noise. They are difficult to sensitively reflect specific pathological changes such as progressive bone resorption of alveolar bone and degeneration of trabecular microstructure. As a result, this information is submerged by imaging noise during feature extraction, which significantly reduces the reliability of CBCT features for alveolar bone mineral density assessment. Summary of the Invention

[0004] To address the technical problem of poor reliability of CBCT features in alveolar bone mineral density assessment due to weakened correlation between morphological parameters and pathological findings, and distortion in the capture of microstructural features, this invention aims to provide a CBCT feature extraction method for alveolar bone mineral density assessment. The specific technical solution adopted is as follows: This invention proposes a CBCT feature extraction method for alveolar bone mineral density assessment, the method comprising: Obtain alveolar bone blocks of the tooth to be analyzed at various time points within the scanning time window; The alveolar bone block is divided into different sub-blocks, and the resorption and degradation significance of the sub-blocks is obtained; based on the difference in the resorption and degradation significance of the different sub-blocks, the sub-blocks in the alveolar bone block are merged to generate different connecting blocks, and disease blocks are selected. The activity of lesions is obtained by considering the clarity of the boundaries of all lesion blocks in the alveolar bone region and the differences in the severity of lesions among different sub-blocks within them. Based on the magnitude and fluctuation of the absorption and degradation significance of sub-blocks in different preset directions of alveolar bone blocks, the spatial dispersion of bone density is obtained; The bone pathological risk level is obtained based on the lesion activity, bone density spatial dispersion, and correlation between the two in all alveolar bone blocks within the scanning time window.

[0005] Further, the acquisition of lesion activity includes: Calculate the variance and range of the absorption and degradation significance of all sub-blocks within each lesion block, and use the product of the variance and the range as the intralesion heterogeneity. The clarity of the lesion boundary is obtained based on the difference in the significance of absorption and degradation between the sub-blocks located at the boundary position within each lesion block and the adjacent sub-blocks within the non-lesion connecting block. The lesion activity is obtained based on the internal heterogeneity of the lesion in all sub-blocks of the alveolar bone block and the clarity of the lesion boundary.

[0006] Furthermore, obtaining the clarity of the lesion boundary includes: The sub-blocks within the connecting blocks of the alveolar bone block, excluding the diseased block, are denoted as non-disease sub-blocks; a first boundary block is selected from the sub-blocks of each diseased block, and at least one of the non-disease sub-blocks exists among the adjacent sub-blocks of the first boundary block; The local clarity of each first boundary block is obtained by averaging the absolute values ​​of the differences in the absorption and degradation significance between each first boundary block and all adjacent non-lesion blocks. The average local sharpness of all first boundary blocks within each lesion block is taken as the boundary sharpness of each lesion block.

[0007] Further, the acquisition of lesion activity includes: The local activity of the corresponding lesion block is obtained based on the internal heterogeneity and the boundary clarity of each lesion block. The ratio of the number of sub-blocks in each lesion block in the alveolar bone block to the total number of sub-blocks in all lesion blocks is recorded as the lesion weight; the local activity of all lesion blocks in the alveolar bone block is weighted and summed according to the lesion weight to obtain the lesion activity of the alveolar bone block.

[0008] Further, obtaining the spatial dispersion of bone mineral density includes: A second boundary block is selected from the sub-blocks of the alveolar bone block, and the edge voxel points of the alveolar bone block are located in the second boundary block; For each sub-block of the alveolar bone block located in each second boundary block in each preset direction, calculate the mean and variance of the significance of absorption and degradation of all sub-blocks respectively, and use the product of the mean and the variance as the local dispersion of each second boundary block in each preset direction. Calculate the mean of the local dispersion of all second boundary blocks in the alveolar bone block in all preset directions, and normalize the mean to obtain the spatial dispersion of bone density of the alveolar bone block.

[0009] Furthermore, the acquisition of bone pathology risk includes: The spatial dispersion of bone density and the lesion activity of alveolar bone blocks at all time points within the scanning time window are arranged in chronological order to obtain spatial discrete sequence and lesion activity sequence, and the correlation coefficient between the spatial discrete sequence and the lesion activity sequence is obtained. The spatial dispersion of bone density and the activity of lesions in the alveolar bone blocks at all time points within the scanning time window are averaged to obtain the overall dispersion and overall activity. By using the correlation coefficient, the overall dispersion and the overall activity are adjusted to obtain the bone pathology risk score.

[0010] Furthermore, the bone pathology risk score is a weighted result of the sum of the overall dispersion and the overall activity, using the sum of the correlation coefficient and the constant 1.

[0011] Furthermore, the generation of different connection blocks and the selection of lesion blocks include: Based on the absolute value of the difference in the significance of absorption and degradation between every two sub-blocks, all sub-blocks in the alveolar bone block are clustered to obtain several clusters; several three-dimensional connected regions are formed by adjacent sub-blocks within the same cluster, denoted as connecting blocks. The average of the resorption and degradation significance of all sub-regions within the alveolar bone block is used to obtain the segmentation threshold; Calculate the mean of the absorption degradation significance of all sub-blocks within each cluster, and identify the connected blocks of clusters whose mean is greater than the partition threshold as lesion blocks.

[0012] Furthermore, the acquisition of the absorption degradation salience of the sub-block includes: Obtain the skewness and kurtosis of the gray values ​​of all voxels within the sub-block; multiply the absolute value of the difference between the kurtosis and the preset normal kurtosis and the skewness as the absorption degradation saliency of the sub-block.

[0013] Furthermore, the heterogeneity within the lesion and the clarity of the lesion boundary are both positively correlated with the local activity.

[0014] The present invention has the following beneficial effects: In this embodiment of the invention, the significance of resorption and degradation, lesion activity, and spatial dispersion of bone density are independent of easily distorted microstructures. They integrate the block continuity, boundary features, internal heterogeneity, and spatial distribution patterns of lesions, extracting biological features highly correlated with bone resorption activity and the degree of microstructural degradation from CBCT images. This reduces distortion in microstructural feature capture and significantly enhances the correlation between CBCT features and the actual pathological state. Lesion activity and spatial dispersion of bone density sequentially represent the degree of bone structural deterioration and the activity of diseased cells in the alveolar bone block. Combining the correlation between the two, the synergy and pattern of lesion destruction are analyzed, overcoming the limitations of traditional morphological parameters and achieving multi-dimensional coupling analysis of pathological progression and spatial heterogeneity, thus improving the reliability of bone pathology risk analysis. This approach elevates the analysis level of CBCT images from morphological observation to pathological function inference, solving the problem of poor reliability in alveolar bone density assessment caused by weakened correlation between morphological parameters and pathology and distortion in microstructural feature capture. Attached Figure Description

[0015] 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.

[0016] Figure 1 The flowchart illustrates the steps of a CBCT feature extraction method for alveolar bone mineral density assessment according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating a method for obtaining lesion activity according to an embodiment of the present invention. Figure 3 This is a schematic diagram of a computer device for CBCT feature extraction for alveolar bone mineral density assessment, provided as an embodiment of the present invention. Detailed Implementation

[0017] 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 CBCT feature extraction method for alveolar bone mineral density assessment 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.

[0018] 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.

[0019] The following describes in detail, with reference to the accompanying drawings, a specific scheme for CBCT feature extraction method for alveolar bone mineral density assessment provided by the present invention.

[0020] Example 1: This invention proposes a CBCT feature extraction method for alveolar bone mineral density assessment. Please refer to [link to relevant documentation]. Figure 1 The diagram illustrates a flowchart of a CBCT feature extraction method for alveolar bone mineral density assessment according to an embodiment of the present invention. The method includes: Step S1: Obtain alveolar bone blocks of the tooth to be analyzed at various time points within the scanning time window.

[0021] Using cone-beam computed tomography (CBCT), alveolar bone CBCT images were acquired at different viewpoints within a unified scanning protocol at each time point within the scanning window. These images completely captured the anatomical information of the tooth to be analyzed. Three-dimensional reconstruction was performed on all alveolar bone CBCT images from the same time point to obtain a three-dimensional jawbone model for each time point. The tooth to be analyzed was determined in the three-dimensional jawbone model, and the three-dimensional volume encompassing the alveolar ridge crest 1 mm radicularly to the root apex 1 mm coronally was used as the alveolar bone block for each time point. Key scanning parameters included: peak voltage of 80 to 90 kV, tube current of 4 to 10 mA, voxel size of 0.08 to 0.25 mm, and scanning field diameter of 4 to 8 cm. To reduce motion artifacts, the patient's head was fixed during scanning using a positioning device to ensure the occlusal plane was parallel to the horizontal plane.

[0022] In one implementation of this invention, the length of the scanning time window is set to 20 seconds, and the time interval between two adjacent time points is set to 2 seconds. The implementer can set these values ​​according to specific circumstances.

[0023] Step S2: Divide the alveolar bone block into different sub-blocks and obtain the significance of resorption and degradation of the sub-blocks; based on the difference in the significance of resorption and degradation of different sub-blocks, merge the sub-blocks in the alveolar bone block to generate different connecting blocks, and select the lesion block.

[0024] Alveolar bone lesions present with complex radiographic manifestations. To accurately analyze abnormal alveolar bone mineral density, the alveolar bone block is divided into sub-blocks. Typically, pathological changes in bone density, such as alveolar bone resorption and microstructural disarray, do not exist in isolation but rather present as continuously distributed lesions. Therefore, it is necessary to merge sub-blocks within the alveolar bone block. The significance of resorption and degradation reflects the degree of bone resorption and microstructural deterioration in the sub-block, presenting the local pathological state of the sub-block. By merging blocks based on the differences in the significance of resorption and degradation among different sub-blocks, connecting blocks representing potential continuous lesions are generated, determining the precise boundaries of the lesions, which is crucial for subsequent lesion assessment. The lesion block represents the alveolar bone mineral density loss focus.

[0025] In one implementation of this invention, the alveolar bone region is uniformly divided into different sub-regions, each sub-region having a size of 3 voxels. 3 voxels 3 voxels.

[0026] Step S3: Obtain the lesion activity based on the clarity of the boundaries of all lesion blocks in the alveolar bone area and the differences in the lesion degree of different sub-blocks within them.

[0027] Significant heterogeneity is often observed within alveolar bone lesions, a characteristic correlated with the stage of disease progression. Specifically, in the active progression phase of lesions, alveolar bone resorption spreads from the local area to the surrounding tissue, resulting in significant differences in the severity of lesions within the lesion. In the stable phase, alveolar bone resorption ceases, and the overall pathological state of the lesion stabilizes, leading to smaller differences in the severity of lesions within the lesion. Furthermore, the boundary between alveolar bone lesions and healthy tissue exhibits a steep transition during the active progression phase, while the boundary of alveolar bone lesions in the stable phase shows a gentle transition. By comprehensively analyzing the differences in lesion severity among different sub-regions within the lesion area of ​​the alveolar bone block and the clarity of the lesion block's boundary, the probability that the corresponding lesion in the lesion block is in the active progression phase can be determined, thus obtaining the lesion activity level.

[0028] Step S4: Obtain the spatial dispersion of bone density based on the magnitude and fluctuation of the absorption and degradation of sub-blocks in different preset directions of the alveolar bone block.

[0029] Pathological changes in alveolar bone, such as bone resorption and microstructural deterioration, often begin locally, forming scattered lesions interwoven with healthy bone. The core spatial characteristics of these lesions exhibit abrupt transitions between healthy tissue and the lesion itself, resulting in high and drastic variations in the degree of resorption and deterioration in sub-blocks across multiple pre-defined directions. However, healthy alveolar bone shows a relatively uniform distribution of bone density across these pre-defined directions, leading to lower and more consistent levels of resorption and deterioration in sub-blocks within each direction. Therefore, the spatial dispersion of bone density can be obtained by comprehensively considering the magnitude and fluctuation of the resorption and deterioration in sub-blocks along pre-defined directions.

[0030] Step S5: Obtain the bone pathological risk level based on the lesion activity, spatial dispersion of bone density, and correlation between the two in all alveolar bone blocks within the scanning time window.

[0031] Lesion activity and spatial dispersion of bone mineral density successively represent the degree of bone structure deterioration and the activity of diseased cells in the alveolar bone block. Combining the two can analyze the basic risk level of alveolar bone mineral density pathology. By combining the correlation between lesion activity and spatial dispersion of bone mineral density, the synergy and pattern of lesion destruction can be distinguished, and different pathological states of alveolar bone mineral density can be achieved, thus realizing a more accurate analysis of bone pathological risk.

[0032] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the absorption degradation saliency includes: obtaining the skewness and kurtosis of the gray values ​​of all voxel points in the sub-block; and using the product of the absolute value of the difference between the kurtosis and the preset normal kurtosis and the skewness as the absorption degradation saliency of the sub-block.

[0033] It should be noted that the core pathological feature of early alveolar bone resorption or loss of bone density is localized mineral loss. The grayscale value of a voxel reflects the degree of X-ray attenuation at that location. The higher the mineral content in the bone tissue, the stronger the X-ray attenuation, and the weaker the signal received by the detector. Therefore, the grayscale value of CBCT images is positively correlated with the mineral content of bone tissue, and changes in grayscale distribution can reflect abnormal bone density caused by early bone resorption.

[0034] The alveolar bone has a regular trabecular arrangement and uniform mineral distribution, with its grayscale distribution approximating a normal distribution. However, bone resorption leads to local mineral loss, resulting in an increase in low-grayscale voxels and a right-skewed distribution. Simultaneously, alveolar bone structural degeneration causes a more dispersed bone density distribution or the formation of abnormal peaks due to microcystic areas, manifested as decreased or increased kurtosis. Therefore, a greater skewness indicates a higher proportion of low-grayscale voxels within the sub-block, and thus a more significant degree of local bone resorption in that sub-block. A greater difference between the kurtosis and the preset normal kurtosis indicates a disordered local microstructure of the alveolar bone. Since the kurtosis of a normal distribution is 3, this embodiment sets the preset normal kurtosis to 3. The greater the significance of resorption and degeneration, the more pronounced the right-skewed grayscale distribution and kurtosis anomalies in the sub-block, and the more significant the bone resorption and microstructural degeneration in that sub-block.

[0035] Preferably, in some possible implementations of the embodiments of the present invention, the method for selecting lesion blocks includes: clustering all sub-blocks in the alveolar bone block based on the absolute value of the difference in the significance of resorption and degradation between every two sub-blocks to obtain several clusters; forming several three-dimensional connected regions from adjacent sub-blocks within the same cluster, denoted as connecting blocks; averaging the significance of resorption and degradation of all sub-blocks in the alveolar bone block to obtain a partitioning threshold; calculating the mean of the significance of resorption and degradation of all sub-blocks within each cluster, and selecting the connecting blocks corresponding to clusters with a mean greater than the partitioning threshold as lesion blocks.

[0036] It should be noted that clustering only considers feature similarity; however, lesions are usually continuous. Constructing connected units introduces spatial continuity constraints, making connected blocks more consistent with the actual lesion morphology. The partitioning threshold is the overall normal level of the absorption and degradation significance of all sub-blocks. Clusters with a mean greater than the partitioning threshold have a higher probability of corresponding connected blocks being lesions, and are thus designated as lesion blocks. In this embodiment of the invention, sub-blocks within a 6-adjacent range of each sub-block are recorded as its adjacent sub-blocks, meaning each sub-block shares a complete surface with its adjacent sub-blocks; alternatively, it can be a sub-region within an 18-adjacent range.

[0037] In one implementation of this invention, the K-means clustering algorithm is used to cluster sub-blocks in the alveolar bone block; wherein the K value is determined by the elbow method; the absolute value of the difference in the significance of absorption and degradation between two sub-blocks is equivalent to the distance.

[0038] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining lesion activity is described in [reference needed]. Figure 2 The diagram illustrates a flowchart of a method for obtaining lesion activity according to an embodiment of the present invention, the method comprising: Step S310: Calculate the variance and range of the absorption and degradation significance of all sub-blocks within each lesion block, and use the product of the variance and range as the heterogeneity within the lesion.

[0039] It should be noted that, because the pathological severity of lesions gradually decreases from the center to the edge during the progression phase of active lesions, while the pathology within lesions tends to stabilize during the stable phase, and the degree of absorption and degradation reflects the local pathological state of sub-blocks, the differences in the degree of absorption and degradation among different sub-blocks within lesions during the progression phase of active lesions are relatively large, while the differences are smaller among different sub-blocks within lesions during the stable phase. A larger variance indicates a more significant difference in the pathological state among different sub-blocks within the lesion area. The range directly reflects the "gradient span" of bone density loss within the lesion and is a key quantitative indicator of the unevenness of lesion progression; a larger range indicates a more significant difference in the degree of pathological severity within the corresponding lesion area. Therefore, both variance and range are positively correlated with intralesional heterogeneity. A greater intralesional heterogeneity indicates a significant pathological gradient within the corresponding lesion area, making it more likely that the lesion is in the progression phase of an active lesion; conversely, a smaller heterogeneity indicates a more uniform degree of lesion severity, making it more likely that the lesion is in the stable phase.

[0040] Step S320: Obtain the lesion boundary clarity based on the difference in the absorption and degradation significance of sub-blocks located at the boundary within each lesion block and adjacent sub-blocks within non-lesion connected blocks.

[0041] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the clarity of the lesion boundary includes: designating sub-blocks within the connecting blocks of the alveolar bone block other than the lesion block as non-lesion sub-blocks; selecting a first boundary block from the sub-blocks of each lesion block, wherein at least one non-lesion sub-block exists among the adjacent sub-blocks of the first boundary block; averaging the absolute values ​​of the differences in the absorption and degradation significance of each first boundary block with all its adjacent non-lesion blocks to obtain the local clarity of each first boundary block; and taking the average of the local clarity of all first boundary blocks within each lesion block as the boundary clarity of each lesion block.

[0042] It should be noted that the boundary between alveolar bone lesions and healthy tissue exhibits a "steep transition" characteristic during the disease progression phase, with the degree of resorption and degradation decreasing rapidly from the diseased tissue to the healthy tissue. In contrast, the boundary of alveolar bone lesions in the stable phase shows a gentle transition, with the degree of resorption and degradation gradually changing in sub-regions at the lesion edge. Local clarity reflects the local boundary gradient of the lesion corresponding to the lesion block. By comprehensively analyzing the local clarity of all boundary sub-regions within the lesion block, the complete boundary gradient of the lesion is obtained, yielding the boundary clarity. The greater the boundary clarity, the greater the heterogeneity difference within the lesion at the boundary of the lesion block, meaning a clearer boundary between the lesion and healthy tissue, and a higher probability that the lesion is in the active disease progression phase.

[0043] Step S330: Obtain the lesion activity based on the internal heterogeneity and boundary clarity of all sub-blocks in the alveolar bone block.

[0044] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining lesion activity includes: obtaining the local activity of the corresponding lesion block based on the lesion heterogeneity and lesion boundary clarity of each lesion block; recording the ratio of the number of sub-blocks in each lesion block in the alveolar bone block to the total number of sub-blocks in all lesion blocks as the lesion weight; and weighting and summing the local activity of all lesion blocks in the alveolar bone block according to the lesion weight to obtain the lesion activity of the alveolar bone block.

[0045] It should be noted that the greater the internal heterogeneity and the clarity of the lesion boundary, the greater the likelihood that the corresponding lesion block is in an active disease progression phase. Therefore, both internal heterogeneity and lesion boundary clarity are positively correlated with local activity. In this embodiment of the invention, the product of the internal heterogeneity and the clarity of the lesion boundary for each lesion block is normalized to obtain the local activity. This embodiment of the invention uses min-max normalization for normalization; however, other normalization methods such as the Sigmoid function, function transformation, and min-max normalization can also be used, and are not limited here.

[0046] In one specific implementation of this invention, the lesion activity is expressed by the formula: In the formula, The activity level of the lesion in the alveolar bone region; K is the total number of lesion regions in the alveolar bone region; This represents the number of sub-blocks within the k-th lesion block in the alveolar bone region; This represents the total number of sub-blocks within all diseased areas in the alveolar bone region; The lesion weight is the weight of the k-th lesion block in the alveolar bone region. This represents the local activity of the k-th lesion block within the alveolar bone region.

[0047] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the spatial dispersion of bone density includes: selecting a second boundary block from the sub-blocks of the alveolar bone block, wherein the edge voxel points of the alveolar bone block are located in the second boundary block; for the sub-blocks in the alveolar bone block located in each second boundary block in each preset direction, calculating the mean and variance of the absorption and degradation significance of all sub-blocks respectively, and using the product of the mean and variance as the local dispersion of each second boundary block in each preset direction; calculating the mean of the local dispersion of all second boundary blocks in the alveolar bone block in each preset direction, and normalizing the mean to obtain the spatial dispersion of bone density of the alveolar bone block.

[0048] It should be noted that for sub-blocks within each second boundary block in each preset direction of the alveolar bone region, the central voxel point of these sub-blocks lies on a straight line passing through the central voxel point of the second boundary block. Healthy alveolar bone exhibits a relatively uniform bone density distribution across preset directions, resulting in lower and more consistent resorption and degradation significance in sub-blocks across these directions. In contrast, diseased alveolar bone contains numerous small lesions, leading to higher and more drastic variations in resorption and degradation significance across sub-blocks in multiple preset directions. Smaller mean and variance indicate less dispersion in the resorption and degradation significance of sub-blocks within the preset direction of the second boundary block, thus reducing local dispersion and increasing the likelihood that the sub-block within the preset direction of the second boundary block belongs to healthy alveolar bone. By averaging the local dispersion of all second boundary blocks in all preset directions within the alveolar bone region, the spatial dispersion of bone density is obtained, which describes the distribution of pathological conditions within the entire alveolar bone region. This directly reflects the overall pathological deterioration of the alveolar bone region. The greater the spatial dispersion of bone density, the more dispersed the micro-lesions within the alveolar bone region, and the more severe the deterioration of the bone structure.

[0049] In one specific implementation of this invention, the spatial dispersion of bone density (SDI) of the alveolar bone block is expressed by the formula: In the formula, M represents the number of second boundary blocks in the alveolar bone region; G represents the number of preset directions. denoted as , where is the local discreteness of the m-th second boundary block in the alveolar bone region along the g-th preset direction; Norm is the normalization function.

[0050] In one implementation of this invention, the preset direction includes the vertical direction from the alveolar ridge crest to the root apex and the mesiodistal horizontal direction.

[0051] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the bone pathology risk includes: arranging the spatial dispersion of bone density and the lesion activity of alveolar bone blocks at all time points within the scanning time window in chronological order to obtain a spatially discrete sequence and a lesion activity sequence, and obtaining the correlation coefficient between the spatially discrete sequence and the lesion activity sequence; averaging the spatial dispersion of bone density and the lesion activity of alveolar bone blocks at all time points within the scanning time window to obtain the overall dispersion and the overall activity; and adjusting the overall dispersion and the overall activity using the correlation coefficient to obtain the bone pathology risk.

[0052] It should be noted that when the correlation coefficient is closer to 1, the spatial dispersion of bone density and the activity of lesions increase synchronously. This means the lesions in the alveolar bone area are more dispersed and the lesions are more active, pathologically corresponding to "widespread active bone resorption." Bone resorption starts simultaneously from multiple sites and is in the progressive stage, with bone density loss exhibiting a "wide range and rapid speed" characteristic, and the risk has a synergistic additive effect. When the correlation coefficient is close to 0, it indicates that the global and local characteristics of the alveolar bone area are asynchronous, and the risk is mainly dominated by a single dimension, without the need for synergistic amplification. When the correlation coefficient is close to -1, it indicates that the increase in spatial dispersion of bone density and the decrease in lesion activity are synchronous. This means the lesions in the alveolar bone area are more dispersed and stable, pathologically corresponding to "multiple small stable lesions." Bone resorption has stopped, leaving only scattered old lesions, and the risk exhibits a decaying effect. The greater the overall dispersion and overall activity, the more severe the bone structure deterioration and the stronger the activity of diseased cells in the alveolar bone block within the scanning time window. This indicates a more severe bone structure and biological destructive force. Both represent the basic risk level of alveolar bone density pathology. By adjusting the correlation coefficient, a more accurate analysis of bone pathology risk can be achieved.

[0053] In this embodiment of the invention, the sum of the correlation coefficient and the constant 1 is used to weight the sum of the overall dispersion and the overall activity to obtain the bone pathology risk score. A higher bone pathology risk score indicates a higher overall risk of alveolar bone density loss and a worse bone health status.

[0054] In one implementation of this invention, the correlation coefficient refers to the Pearson correlation coefficient.

[0055] In this embodiment of the invention, the bone pathological risk of the tooth position to be analyzed is input into a pre-trained neural network. This neural network has been trained using a large amount of labeled clinical data, such as bone density categories labeled as normal bone, mild osteoporosis, and severe osteoporosis, in order to learn the complex mapping relationship between bone pathological risk and bone category. The neural network outputs the bone density classification result of the alveolar bone of the tooth position to be analyzed.

[0056] The classification results of this scheme are based on objective feature annotation results generated by big data analysis. It does not directly provide diagnostic conclusions, but provides doctors with objective quantitative data references, enabling doctors to focus on the pathological characteristics of alveolar bone more quickly and improve the efficiency of doctors in recognizing alveolar bone mineral density classification results. The final medical judgment still needs to be made by professional doctors.

[0057] This invention is now complete.

[0058] Example 2: This invention also presents a schematic diagram of a computer device for CBCT feature extraction in alveolar bone mineral density assessment; please refer to [link / reference]. Figure 3The computer device includes a memory 601, a processor 602, and a computer program 603 stored in the memory 601 and running on the processor 602. When the processor 602 executes the computer program 603, the computer device can perform any of the CBCT feature extraction methods for alveolar bone mineral density assessment described above.

[0059] 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 a CBCT feature extraction method for alveolar bone mineral density assessment provided in embodiments of this application.

[0060] 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.

[0061] When each module is divided according to its function, the device may also include a communication module, a signal analysis module, a complexity analysis module, and a positioning module. It should be noted that all relevant content of each step involved in the above method embodiments can be referenced from the functional descriptions of the corresponding functional modules, and will not be repeated here.

[0062] It should be understood that the device provided in this embodiment is used to perform the above-described CBCT feature extraction method for alveolar bone mineral density assessment, and therefore can achieve the same effect as the above-described implementation method.

[0063] 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.

[0064] 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 a microprocessor, etc., and the storage module may be a memory.

[0065] Example 3: This embodiment also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer, the computer executes the above-described related method steps to implement the CBCT feature extraction method for alveolar bone mineral density assessment provided in the above embodiment.

[0066] The apparatus and computer-readable storage medium provided in this embodiment are used to execute the corresponding methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0067] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0068] 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.

[0069] 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.

Claims

1. A CBCT feature extraction method for alveolar bone mineral density assessment, characterized in that, The method includes: Obtain alveolar bone blocks of the tooth to be analyzed at various time points within the scanning time window; The alveolar bone block is divided into different sub-blocks, and the resorption and degradation significance of the sub-blocks is obtained; based on the difference in the resorption and degradation significance of the different sub-blocks, the sub-blocks in the alveolar bone block are merged to generate different connecting blocks, and disease blocks are selected. The activity of lesions is obtained by considering the clarity of the boundaries of all lesion blocks in the alveolar bone region and the differences in the severity of lesions among different sub-blocks within them. Based on the magnitude and fluctuation of the absorption and degradation significance of sub-blocks in different preset directions of alveolar bone blocks, the spatial dispersion of bone density is obtained; The bone pathological risk level is obtained based on the lesion activity, bone density spatial dispersion, and correlation between the two in all alveolar bone blocks within the scanning time window.

2. The CBCT feature extraction method for alveolar bone mineral density assessment according to claim 1, characterized in that, The acquisition of lesion activity includes: Calculate the variance and range of the absorption and degradation significance of all sub-blocks within each lesion block, and use the product of the variance and the range as the intralesion heterogeneity. The clarity of the lesion boundary is obtained based on the difference in the significance of absorption and degradation between the sub-blocks located at the boundary position within each lesion block and the adjacent sub-blocks within the non-lesion connecting block. The lesion activity is obtained based on the internal heterogeneity of the lesion in all sub-blocks of the alveolar bone block and the clarity of the lesion boundary.

3. The CBCT feature extraction method for alveolar bone mineral density assessment according to claim 2, characterized in that, The acquisition of lesion boundary clarity includes: The sub-blocks within the connecting blocks of the alveolar bone block, excluding the diseased block, are denoted as non-disease sub-blocks; a first boundary block is selected from the sub-blocks of each diseased block, and at least one of the non-disease sub-blocks exists among the adjacent sub-blocks of the first boundary block; The local clarity of each first boundary block is obtained by averaging the absolute values ​​of the differences in the absorption and degradation significance between each first boundary block and all adjacent non-lesion blocks. The average local sharpness of all first boundary blocks within each lesion block is taken as the boundary sharpness of each lesion block.

4. The CBCT feature extraction method for alveolar bone mineral density assessment according to claim 2, characterized in that, The acquisition of lesion activity includes: The local activity of the corresponding lesion block is obtained based on the internal heterogeneity and the boundary clarity of each lesion block. The ratio of the number of sub-blocks in each lesion block in the alveolar bone block to the total number of sub-blocks in all lesion blocks is recorded as the lesion weight; the local activity of all lesion blocks in the alveolar bone block is weighted and summed according to the lesion weight to obtain the lesion activity of the alveolar bone block.

5. The CBCT feature extraction method for alveolar bone mineral density assessment according to claim 1, characterized in that, The acquisition of bone mineral density spatial dispersion includes: A second boundary block is selected from the sub-blocks of the alveolar bone block, and the edge voxel points of the alveolar bone block are located in the second boundary block; For each sub-block of the alveolar bone block located in each second boundary block in each preset direction, calculate the mean and variance of the significance of absorption and degradation of all sub-blocks respectively, and use the product of the mean and the variance as the local dispersion of each second boundary block in each preset direction. Calculate the mean of the local dispersion of all second boundary blocks in the alveolar bone block in all preset directions, and normalize the mean to obtain the spatial dispersion of bone density of the alveolar bone block.

6. The CBCT feature extraction method for alveolar bone mineral density assessment according to claim 1, characterized in that, The acquisition of bone pathological risk includes: The spatial dispersion of bone density and the lesion activity of alveolar bone blocks at all time points within the scanning time window are arranged in chronological order to obtain spatial discrete sequence and lesion activity sequence, and the correlation coefficient between the spatial discrete sequence and the lesion activity sequence is obtained. The spatial dispersion of bone density and the activity of lesions in the alveolar bone blocks at all time points within the scanning time window are averaged to obtain the overall dispersion and overall activity. By using the correlation coefficient, the overall dispersion and the overall activity are adjusted to obtain the bone pathology risk score.

7. The CBCT feature extraction method for alveolar bone mineral density assessment according to claim 6, characterized in that, The bone pathological risk score is a weighted result of the sum of the overall dispersion and the overall activity, using the sum of the correlation coefficient and the constant 1.

8. The CBCT feature extraction method for alveolar bone mineral density assessment according to claim 1, characterized in that, The generation of different connectivity blocks and the selection of lesion blocks include: Based on the absolute value of the difference in the significance of absorption and degradation between every two sub-blocks, all sub-blocks in the alveolar bone block are clustered to obtain several clusters; several three-dimensional connected regions are formed by adjacent sub-blocks within the same cluster, denoted as connecting blocks. The average of the resorption and degradation significance of all sub-regions within the alveolar bone block is used to obtain the segmentation threshold; Calculate the mean of the absorption degradation significance of all sub-blocks within each cluster, and identify the connected blocks of clusters whose mean is greater than the partition threshold as lesion blocks.

9. The CBCT feature extraction method for alveolar bone mineral density assessment according to claim 1, characterized in that, The acquisition of the absorption degradation significance of the sub-block includes: Obtain the skewness and kurtosis of the gray values ​​of all voxels within the sub-block; multiply the absolute value of the difference between the kurtosis and the preset normal kurtosis and the skewness as the absorption degradation saliency of the sub-block.

10. A CBCT feature extraction method for alveolar bone mineral density assessment according to claim 4, characterized in that, The heterogeneity within the lesion and the clarity of the lesion boundary are both positively correlated with the local activity.

Images