An intelligent processing method for oral and maxillofacial CBCT images

Abstract

The present invention relates to the field of image enhancement technology, and more specifically to an intelligent processing method for oral and maxillofacial CBCT images, comprising: determining an enhancement coefficient corresponding to each pixel in the image to be processed based on the grayscale difference value before and after histogram equalization, and the difference value of the degree of visibility of each pixel relative to neighboring pixels before and after histogram equalization; determining a final grayscale change corresponding to each grayscale level based on the enhancement coefficient corresponding to each pixel; and determining a final grayscale level for each pixel based on the current grayscale level of each pixel and the final grayscale change corresponding to each grayscale level, thereby enhancing the image to be processed. This method has a good image enhancement effect.

Description

Technical Field

[0001] The present invention relates to the technical field of image enhancement, and in particular to an intelligent processing method for oral and maxillofacial CBCT images. Background Art

[0002] In the field of medical imaging, the diagnosis and treatment of patients with oral caries are crucial. Cone beam CT (CBCT) is a widely used and effective method for detecting and analyzing oral caries, providing detailed information about the dental structure, morphology, and alveolar bone condition. However, CBCT images are often complex and noisy, significantly interfering with clinicians' accurate caries assessment. Furthermore, because different caries cycles correspond to different caries, early caries assessments can lead to misjudgments.

[0003] Furthermore, oral and maxillofacial CBCT image enhancement is currently performed using a histogram equalization method. However, when using histogram equalization to perform oral and maxillofacial CBCT image enhancement, the enhancement effect may be poor due to the lack of obvious dental caries. Summary of the Invention

[0004] An intelligent processing method for oral and maxillofacial CBCT images has a good enhancement effect and prevents the phenomenon of poor enhancement effect when the caries area is not obvious.

[0005] According to the grayscale difference of each pixel point in the neighborhood range before and after grayscale histogram equalization in the image to be processed, data matching is performed to obtain the distance matrix and the optimal distance path. According to the overall situation of the minimum distance between each distance value on the optimal distance path and the data on the diagonal line of the distance matrix, as well as all distance values ​​on the optimal distance path, the effect evaluation value of each pixel point in the image to be processed is obtained;

[0006] Determine the enhancement coefficient corresponding to each pixel point based on the grayscale difference value of each pixel point in the image to be processed before and after histogram equalization and the effect evaluation value of each pixel point;

[0007] Determine the final grayscale change corresponding to each grayscale based on the enhancement coefficient corresponding to each pixel, the probability distribution of each pixel belonging to the oral caries area, and the grayscale change before and after histogram equalization;

[0008] The final grayscale of each pixel is determined based on the current grayscale of each pixel and the final grayscale variation corresponding to each grayscale, thereby performing image enhancement on the image to be processed.

[0009] Preferably, performing data matching based on the grayscale difference of each pixel in the image to be processed within the neighborhood range before and after grayscale histogram equalization to obtain the distance matrix and the optimal distance path includes:

[0010] Any pixel in the image to be processed is recorded as the target pixel, and each pixel in the neighborhood of the target pixel is recorded as a reference pixel;

[0011] Before the image to be processed is subjected to histogram equalization, the difference in grayscale value between each reference pixel and the target pixel is obtained to form a first grayscale difference sequence; after the image to be processed is subjected to histogram equalization, the difference in grayscale value between each reference pixel and the target pixel is obtained to form a second grayscale difference sequence;

[0012] A dynamic time warping algorithm is used to obtain a distance matrix and an optimal distance path when matching the first grayscale difference sequence and the second grayscale difference sequence; each element in the distance matrix is ​​a distance value.

[0013] Preferably, the method of obtaining the effect evaluation value of each pixel in the image to be processed based on the overall situation of the minimum distance between each distance value on the optimal distance path and the data on the diagonal line of the distance matrix, as well as all distance values ​​on the optimal distance path, includes:

[0014] Record any distance value on the optimal distance path as the selected distance value, calculate the minimum value of the Euclidean distance between the selected distance value and each element on the diagonal of the distance matrix, and record it as the distance difference eigenvalue of the selected distance value; record the cumulative sum of the distance difference eigenvalues ​​of all distance values ​​on the optimal distance path as the first cumulative value;

[0015] The cumulative sum of all distance values ​​on the optimal distance path is recorded as the second cumulative value; the effect evaluation value of the target pixel point is obtained according to the first cumulative value and the second cumulative value, and the first cumulative value and the effect evaluation value are negatively correlated, and the second cumulative value and the effect evaluation value are positively correlated.

[0016] Preferably, the enhancement coefficient corresponding to the pixel point is calculated as follows:

[0017]

[0018] in, is the enhancement coefficient of the i-th pixel, is the grayscale difference value of the i-th pixel before and after histogram equalization, is the effect evaluation value of the i-th pixel relative to the neighboring pixels before and after histogram equalization, is an exponential function of a natural constant.

[0019] Preferably, the final grayscale change corresponding to each grayscale level is determined based on the enhancement coefficient corresponding to each pixel point, the probability distribution of each pixel point belonging to the oral caries area, and the grayscale change before and after histogram equalization, including:

[0020] The image to be processed before histogram equalization is processed using a neural network to obtain a Gaussian heat map, and the probability of each pixel belonging to oral caries is obtained based on the Gaussian heat map:

[0021] Calculate the average value of the probability that all pixels corresponding to each gray level belong to oral caries to obtain the average value of the probability that the pixel points corresponding to each gray level belong to oral caries;

[0022] The final grayscale change corresponding to each grayscale level is determined based on the probability mean and the enhancement coefficient mean of the pixel points corresponding to each grayscale level.

[0023] Preferably, determining the final grayscale change corresponding to each grayscale based on the probability mean and the mean value of the enhancement coefficients of the pixels corresponding to each grayscale includes:

[0024] Determine the initial gray level after histogram equalization of each gray level;

[0025] The final grayscale change corresponding to each grayscale level is determined by using the initial grayscale change between the initial grayscale level and the current grayscale level, the probability mean, and the enhancement coefficient mean of the pixel points corresponding to each grayscale level.

[0026] Preferably, the final grayscale change corresponding to each grayscale level is calculated as follows:

[0027]

[0028] in, Indicates the final grayscale change corresponding to the jth grayscale level, represents the initial gray level after the j-th gray level is enhanced, It represents the normalized value of the probability mean of the pixel corresponding to the jth gray level belonging to oral caries, represents the value after mean normalization of the enhancement coefficient, is an exponential function of a natural constant.

[0029] Preferably, determining the final grayscale level of each pixel based on the current grayscale level of each pixel and the final grayscale level change corresponding to each grayscale level, thereby performing image enhancement on the image to be processed, includes:

[0030] The current grayscale level of each pixel point is added to the final grayscale level change corresponding to each grayscale level, the addition result is normalized, the normalized result is multiplied by 255 and rounded down to obtain the final grayscale level of each pixel point, thereby performing image enhancement on the image to be processed.

[0031] The embodiments of the present invention have at least the following beneficial effects:

[0032] The present invention first evaluates the effect of preliminary image enhancement by histogram equalization based on the grayscale changes of pixels in a neighborhood before and after histogram equalization of the image to be processed, obtaining an effect evaluation value; then determines the enhancement coefficient corresponding to each pixel based on the grayscale difference value of each pixel in the image to be processed before and after histogram equalization and the effect evaluation value; then determines the final grayscale change corresponding to each grayscale level based on the enhancement coefficient corresponding to each pixel; and finally determines the final grayscale level of each pixel based on the current grayscale level of each pixel and the final grayscale change corresponding to each grayscale level, thereby performing image enhancement on the image to be processed. The method obtains the enhancement effect of each pixel in the current image by comparing the degree of change before and after image enhancement, and then uses a neural network to obtain the probability of each pixel belonging to dental caries, and optimizes and adjusts the enhancement effect of the corresponding grayscale level of each pixel, so that areas with a high probability of dental caries have a better enhancement effect, and prevents the enhancement effect from being unclear when dental caries areas are less obvious. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0034] Figure 1 FIG. 1 is a flow chart of an embodiment of an intelligent processing method for oral and maxillofacial CBCT images according to the present invention. DETAILED DESCRIPTION

[0035] To further illustrate the technical means and effectiveness of the present invention in achieving its intended objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of an intelligent processing method for oral and maxillofacial CBCT images proposed in accordance with the present invention. In the following description, references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0036] Unless defined otherwise, 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 belongs.

[0037] The specific scheme of the intelligent processing method of oral and maxillofacial CBCT images provided by the present invention is described in detail below with reference to the accompanying drawings.

[0038] See also Figure 1 , which shows a flowchart of a method for intelligent processing of oral and maxillofacial CBCT images provided by one embodiment of the present invention, the method comprising the following steps:

[0039] Step S11: performing data matching based on the grayscale difference of each pixel point in the image to be processed in the neighborhood range before and after grayscale histogram equalization to obtain a distance matrix and an optimal distance path; obtaining an effect evaluation value of each pixel point in the image to be processed based on the overall situation of the minimum distance between each distance value on the optimal distance path and the data on the diagonal line of the distance matrix, as well as all distance values ​​on the optimal distance path; determining the enhancement coefficient corresponding to each pixel point based on the grayscale difference value of each pixel point in the image to be processed before and after histogram equalization and the effect evaluation value of each pixel point.

[0040] This application needs to obtain oral and maxillofacial CBCT image data, which is usually stored in DICOM format. In order to facilitate optimization processing, it is converted into a two-dimensional array matrix to facilitate subsequent processing.

[0041] When collecting data, the available equipment includes a CBCT scanner, server, and network transmission system, which can be adjusted by the implementer based on the specific implementation scenario. The CBCT scanner captures the patient's oral and maxillofacial CBCT images; the server stores and manages the raw scan data; and the network transmission system transmits the required data to the workstation for analysis and processing.

[0042] The collected data primarily consists of oral and maxillofacial CBCT images from patients. Each slice image in the oral and maxillofacial CBCT images represents a slice of the oral cavity and can be used to observe dental caries at different locations. This data provides us with rich information for accurate caries identification and extraction during subsequent optimization.

[0043] After acquiring oral and maxillofacial CBCT images at the workstation, a median filter algorithm is used to denoise each slice image to produce the processed image. For this solution, the median filter size is 3*3. The implementer can adjust the median filter size or use other denoising operations based on the specific implementation scenario.

[0044] Specifically, after completing the preprocessing of the oral and maxillofacial CBCT images to obtain the image to be processed, the histogram equalization algorithm can be used to enhance the image to be processed, and then the enhancement coefficient corresponding to each grayscale is obtained based on the change in each grayscale. The area with a higher enhancement coefficient should be the area where caries may exist, and the enhancement coefficient should be lower for the area with a lower possibility of caries. After obtaining the enhancement coefficient corresponding to each pixel point in advance, the enhancement coefficient of each pixel point is adjusted according to the possibility of each pixel point belonging to caries. Based on the enhancement coefficient of each pixel point, the enhancement coefficient of each grayscale level is obtained. Then, based on the enhancement coefficient corresponding to each grayscale level, the cumulative probability distribution function in the histogram equalization is optimized and adjusted, thereby completing the final image enhancement.

[0045] Because when using the histogram equalization algorithm to enhance the image to be processed, the enhancement effect of the histogram equalization algorithm may not guarantee that the area where caries may exist is the best enhancement area, resulting in poor enhancement effect of oral and maxillofacial CBCT images. In order to make the area where caries may exist the best enhancement area, this solution chooses to first obtain the enhancement coefficient corresponding to each pixel point, and according to the initial enhancement coefficient, complete the optimization of the enhancement effect when using histogram equalization for image enhancement. In a specific embodiment, the enhancement coefficient of the i-th pixel point is for:

[0046]

[0047] in, is the enhancement coefficient of the i-th pixel, is the grayscale difference value of the i-th pixel before and after histogram equalization, is the effect evaluation value of the i-th pixel relative to the neighboring pixels before and after histogram equalization, is an exponential function of a natural constant.

[0048] is the grayscale difference value of the i-th pixel before and after histogram equalization. Specifically, the grayscale value of the i-th pixel before histogram equalization is subtracted from the grayscale value of the i-th pixel after histogram equalization to obtain the absolute value of the grayscale difference, thereby obtaining the grayscale difference value of the i-th pixel before and after histogram equalization. The larger the grayscale difference value, the greater the difference in grayscale value of the i-th pixel before and after histogram equalization, and the greater the corresponding enhancement degree, and thus the enhancement coefficient corresponding to the i-th pixel. The larger the value of .

[0049] is the effect evaluation value of the i-th pixel relative to the neighboring pixels before and after histogram equalization. The larger the value is, the greater the change in the visibility of the i-th pixel point. If the i-th pixel point becomes more obvious or less obvious, the better the enhancement effect of the i-th pixel point is. A small value indicates that the visibility of the i-th pixel has a small change, which means that the enhancement effect is not good.

[0050] Obtain the neighborhood pixels of the i-th pixel, for example, a sequence of pixels in an 8-neighborhood region surrounding the i-th pixel. The size of the neighborhood can be adjusted by the implementer based on the specific implementation scenario. Obtain the prominence of the i-th pixel before histogram equalization. The greater the prominence of the i-th pixel, the more prominent it is relative to its neighboring pixels, thereby measuring the prominence of the grayscale of the current i-th pixel relative to the grayscale of its neighboring pixels.

[0051] The difference in grayscale values ​​reflects the prominence of the current pixel. A pixel with a larger grayscale difference than its surrounding pixels will have a higher prominence. The change in prominence of the i-th pixel can be obtained by analyzing the change in grayscale differences within the neighborhood of each pixel in the image before and after grayscale histogram equalization, as well as the distribution of grayscale difference distances.

[0052] Specifically, any pixel in the image to be processed is recorded as a target pixel, and each pixel in the neighborhood of the target pixel is recorded as a reference pixel; before the image to be processed is subjected to histogram equalization, the difference in grayscale value between each reference pixel and the target pixel is obtained to form a first grayscale difference sequence; after the image to be processed is subjected to histogram equalization, the difference in grayscale value between each reference pixel and the target pixel is obtained to form a second grayscale difference sequence.

[0053] After obtaining the difference sequences of the target pixel before and after histogram equalization, the distance matrix algorithm within the DTW algorithm is used to calculate the distance matrix and the corresponding optimal distance path between the two difference sequences. Specifically, the dynamic time warping algorithm is used to obtain the distance matrix and optimal distance path for matching the first grayscale difference sequence and the second grayscale difference sequence; each element in the distance matrix represents a distance value.

[0054] Among them, if the optimal distance path is the diagonal of the distance matrix, it means that the relative difference in the grayscale value of each reference pixel, that is, the pixel points in the neighborhood, before and after enhancement does not change much, and the image is not distorted. If the corresponding optimal distance cumulative value in the two sequences is large at this time, it means that the enhancement effect is good. If the optimal distance path deviates more from the diagonal of the distance matrix, it means that its grayscale values ​​relative to some reference pixels, that is, the pixel points in the neighborhood, are relatively similar, and the grayscale values ​​relative to some neighborhood pixels are significantly different, then the enhancement effect is not good.

[0055] Based on this, the effect evaluation value of the target pixel in the processed image is obtained based on the corresponding relationship and distance distribution between the first and second grayscale difference sequences when analyzed by the dynamic time warping algorithm. In other words, the initial enhancement effect of the pixel is quantitatively evaluated based on the overall minimum distance between each distance value on the optimal distance path and the data on the diagonal of the distance matrix, as well as all distance values ​​on the optimal distance path.

[0056] Record any distance value on the optimal distance path as the selected distance value, calculate the minimum value of the Euclidean distance between the selected distance value and each element on the diagonal of the distance matrix, and record it as the distance difference eigenvalue of the selected distance value; record the cumulative sum of the distance difference eigenvalues ​​of all distance values ​​on the optimal distance path as the first cumulative value ; Record the sum of all distance values ​​on the optimal distance path as the second cumulative value .

[0057] An effect evaluation value of the target pixel point is obtained according to the first accumulated value and the second accumulated value, wherein the first accumulated value and the effect evaluation value are negatively correlated, and the second accumulated value and the effect evaluation value are positively correlated.

[0058] First accumulated value The larger the value is, the more serious the optimal distance value is offset from the diagonal, indicating that the current histogram equalization effect is not good. Therefore, it uses exp(-x) for negative correlation mapping. The second accumulated value The larger the value is, the greater the change before and after the current histogram equalization.

[0059] Based on the first accumulated value and the second accumulated value, the effect evaluation value of the current pixel relative to the neighboring pixel before and after histogram equalization is determined. ,in, It represents the effect evaluation value of the i-th pixel relative to the neighboring pixels before and after histogram equalization. The larger the value, the better the enhancement effect without distortion. or The value of x may be zero, so this solution chooses to use the exp(x) function for mapping. The effect evaluation value of each pixel in the image to be processed represents the effect evaluation of the preliminary image enhancement by histogram equalization.

[0060] Step S12: determining the final grayscale change corresponding to each grayscale based on the enhancement coefficient corresponding to each pixel, the probability distribution of each pixel belonging to the oral caries area, and the grayscale change before and after histogram equalization.

[0061] Specifically, a neural network is used to process the image to be processed before histogram equalization to obtain a Gaussian heat map, and based on the Gaussian heat map, the probability of each pixel belonging to oral caries is obtained. Specifically, a Unet full convolutional network + heatmap thermal map (Gaussian heat map) regression can be used to identify the caries area of ​​the image to be processed, thereby obtaining a Gaussian heat map. And based on the Gaussian heat map, the probability of each pixel belonging to oral caries is obtained. It should be noted that the Gaussian heat map is an image representing probability, and the probability of each pixel belonging to oral caries can be determined by obtaining the Gaussian heat map.

[0062] Based on the probability of each pixel belonging to oral caries, the mean probability of the pixel corresponding to each gray level belonging to oral caries is determined, that is, the mean probability of all pixels corresponding to each gray level belonging to oral caries is calculated to obtain the mean probability of the pixel corresponding to each gray level belonging to oral caries. Specifically, the probability mean of the pixel corresponding to the jth gray level before histogram equalization is obtained. ,in The larger the value, the more pixels corresponding to the j-th gray level belong to oral caries, and the higher the probability.

[0063] The final grayscale change corresponding to each grayscale is determined based on the probability mean and the mean enhancement coefficient of the pixel points corresponding to each grayscale level. Specifically, the mean enhancement coefficient of the enhancement coefficient Z of the pixel points corresponding to the jth grayscale level before and after the enhancement is calculated. , indicating that the j-th gray level has a better enhancement effect before and after enhancement.

[0064] The final grayscale change corresponding to each grayscale level is determined based on the probability mean and the mean enhancement coefficient of the pixel points corresponding to each grayscale level. Specifically, the initial grayscale level after histogram equalization is determined for each grayscale level; the final grayscale change corresponding to each grayscale level is determined using the initial grayscale change between the initial grayscale level and the current grayscale level, the probability mean, and the mean enhancement coefficient of the pixel points corresponding to each grayscale level.

[0065] Specifically, in order to obtain better oral and maxillofacial CBCT image enhancement effect, the gray level with high probability of dental caries should have a higher enhancement effect, and then the gray level corresponding to all gray levels should have a higher enhancement effect. Normalize , and then all gray levels corresponding to Normalize , get the initial gray level of the jth gray level after enhancement , get the initial gray level change , and then the gray level change and and Multiplying them, we can get the new gray level change corresponding to j. :

[0066]

[0067] in is the mean probability that the pixel corresponding to the jth gray level belongs to oral caries The larger the normalized value is, the higher the probability that the current j-th gray level belongs to caries is, and the more obvious the enhancement effect should be.

[0068] is the mean value of the enhancement coefficient corresponding to the jth gray level After normalization, the larger the value is, the better the enhancement effect of the current j-th gray level is.

[0069] and then The larger the value, the higher the probability of dental caries and the better the enhancement effect, so the original grayscale is kept; if the probability of dental caries is higher and the enhancement effect is not good, the grayscale is kept. It can make, compared with the use of , with a greater degree of retention. Currently, if the probability of dental caries is lower, the enhancement effect will be lowered. In order to make the enhancement effect corresponding to the high dental caries probability better, this scheme chooses to use square Then stretch it to get , among which The effect of tensile reinforcement can be selected by the implementer according to the specific implementation scenario.

[0070] Step S13: determining the final grayscale of each pixel based on the current grayscale of each pixel and the final grayscale variation corresponding to each grayscale, thereby performing image enhancement on the image to be processed.

[0071] The current gray level of each pixel is added to the final gray level change corresponding to each gray level, the addition result is normalized, the normalized result is multiplied by 255 and rounded down to obtain the final gray level of each pixel, thereby performing image enhancement on the image to be processed. The normalization and multiplication by 255 are because The value range is .

[0072] The advantage of the present application over the prior art is that the enhancement effect of each pixel in the current image is obtained by measuring the degree of change before and after image enhancement, and then the possibility of each pixel belonging to dental caries is obtained by using a neural network, and the enhancement effect of the corresponding grayscale of each pixel is optimized and adjusted, so that the area with a high possibility of belonging to dental caries has a better enhancement effect, and the enhancement effect is prevented from being unclear when the dental caries area is not obvious.

[0073] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. An intelligent processing method for oral and maxillofacial CBCT images, characterized in that: include: According to the grayscale difference of each pixel point in the neighborhood range before and after grayscale histogram equalization in the image to be processed, data matching is performed to obtain the distance matrix and the optimal distance path. According to the overall situation of the minimum distance between each distance value on the optimal distance path and the data on the diagonal line of the distance matrix, as well as all distance values ​​on the optimal distance path, the effect evaluation value of each pixel point in the image to be processed is obtained; Determine the enhancement coefficient corresponding to each pixel point based on the grayscale difference value of each pixel point in the image to be processed before and after histogram equalization and the effect evaluation value of each pixel point; Determine the final grayscale change corresponding to each grayscale based on the enhancement coefficient corresponding to each pixel, the probability distribution of each pixel belonging to the oral caries area, and the grayscale change before and after histogram equalization; Determining a final grayscale level of each pixel based on a current grayscale level of each pixel and a final grayscale level variation corresponding to each grayscale level, thereby performing image enhancement on the image to be processed; The final grayscale change corresponding to each grayscale is determined based on the enhancement coefficient corresponding to each pixel, the probability distribution of each pixel belonging to the oral caries area, and the grayscale change before and after histogram equalization. include: The image to be processed before histogram equalization is processed using a neural network to obtain a Gaussian heat map, and the probability of each pixel belonging to oral caries is obtained based on the Gaussian heat map: Calculate the average value of the probability that all pixel points corresponding to each gray level belong to oral caries to obtain the average value of the probability that the pixel points corresponding to each gray level belong to oral caries; Determine the final grayscale change corresponding to each grayscale based on the probability mean and the enhancement coefficient mean of the pixel points corresponding to each grayscale; Determining the final grayscale change corresponding to each grayscale based on the probability mean and the enhancement coefficient mean of the pixel points corresponding to each grayscale, including: Determine the initial gray level after histogram equalization of each gray level; Determine the final grayscale change corresponding to each grayscale level using the initial grayscale change between the initial grayscale level and the current grayscale level, the probability mean, and the mean of the enhancement coefficients of the pixels corresponding to each grayscale level; The calculation method of the final grayscale change corresponding to each grayscale is: in, Indicates the final grayscale change corresponding to the j-th grayscale level, represents the initial gray level after the j-th gray level is enhanced, It represents the normalized value of the probability mean of the pixel corresponding to the jth gray level belonging to oral caries, represents the value after mean normalization of the enhancement coefficient, is an exponential function of a natural constant.

2. The intelligent processing method for oral and maxillofacial CBCT images according to claim 1, characterized in that: The method of performing data matching based on the grayscale difference of each pixel in the image to be processed within the neighborhood range before and after grayscale histogram equalization to obtain a distance matrix and an optimal distance path includes: Any pixel in the image to be processed is recorded as the target pixel, and each pixel in the neighborhood of the target pixel is recorded as a reference pixel; Before the image to be processed is subjected to histogram equalization, the difference in grayscale value between each reference pixel and the target pixel is obtained to form a first grayscale difference sequence; after the image to be processed is subjected to histogram equalization, the difference in grayscale value between each reference pixel and the target pixel is obtained to form a second grayscale difference sequence; A dynamic time warping algorithm is used to obtain a distance matrix and an optimal distance path when matching the first grayscale difference sequence and the second grayscale difference sequence; each element in the distance matrix is ​​a distance value.

3. The intelligent processing method for oral and maxillofacial CBCT images according to claim 2, characterized in that: The effect evaluation value of each pixel in the image to be processed is obtained based on the overall situation of the minimum distance between each distance value on the optimal distance path and the data on the diagonal line of the distance matrix, as well as all distance values ​​on the optimal distance path, including: Record any distance value on the optimal distance path as the selected distance value, calculate the minimum value of the Euclidean distance between the selected distance value and each element on the diagonal of the distance matrix, and record it as the distance difference eigenvalue of the selected distance value; record the cumulative sum of the distance difference eigenvalues ​​of all distance values ​​on the optimal distance path as the first cumulative value; The cumulative sum of all distance values ​​on the optimal distance path is recorded as a second cumulative value; an effect evaluation value of the target pixel point is obtained according to the first cumulative value and the second cumulative value, wherein the first cumulative value and the effect evaluation value are negatively correlated, and the second cumulative value and the effect evaluation value are positively correlated; The formula corresponding to the effect evaluation value is: in, Represents the effect evaluation value of the i-th pixel relative to the neighboring pixels before and after histogram equalization; is the first accumulated value; is the second accumulated value.

4. The intelligent processing method for oral and maxillofacial CBCT images according to claim 1, characterized in that: The calculation method of the enhancement coefficient corresponding to the pixel point is: in, is the enhancement coefficient of the i-th pixel, is the grayscale difference value of the i-th pixel before and after histogram equalization, is the effect evaluation value of the i-th pixel relative to the neighboring pixels before and after histogram equalization, is an exponential function of a natural constant.

5. The intelligent processing method for oral and maxillofacial CBCT images according to claim 1, characterized in that: Determining a final grayscale level of each pixel based on a current grayscale level of each pixel and a final grayscale level variation corresponding to each grayscale level, thereby performing image enhancement on the image to be processed, including: The current grayscale level of each pixel point is added to the final grayscale level change corresponding to each grayscale level, the addition result is normalized, the normalized result is multiplied by 255 and rounded down to obtain the final grayscale level of each pixel point, thereby performing image enhancement on the image to be processed.

Images