Intelligent image segmentation system for lower limb trauma fracture detection

Abstract

The invention relates to the technical field of local image enhancement, in particular to an intelligent image segmentation system for lower limb trauma fracture detection. According to the method, the change degree of each pixel point is obtained according to the change rule condition of the pixel points in the lower limb CT image in rows and columns; obtaining the anomaly of each pixel point according to the fluctuation influence condition of the change degree of each pixel point in the row where each pixel point is located; the method comprises the steps of obtaining a preset neighborhood region corresponding to each pixel point, obtaining an adjacent region of the preset neighborhood region corresponding to each pixel point, obtaining the trauma possibility of each pixel point according to the abnormal difference condition between the preset neighborhood region corresponding to each pixel point and each adjacent region and the symmetry of the adjacent regions, obtaining a lower limb enhanced CT image, and segmenting the lower limb enhanced CT image. According to the method, a more accurate and clearer image segmentation result is obtained by enhancing the pixel point contrast.

Description

Technical Field

[0001] The present invention relates to the technical field of local image enhancement, and particularly relates to an intelligent image segmentation system for detecting lower limb trauma fractures. Background Art

[0002] Medical CT images are a commonly used imaging technology for obtaining high-resolution three-dimensional images of the human body. The main purpose of medical CT image segmentation is to extract the tissue structures or specific regions in the CT images and segment them into different regions or objects. Therefore, image segmentation belongs to a relatively important image processing technology.

[0003] When segmenting lower limb CT images in the prior art, the situation that the pixel values of the pixel points in the trauma fracture region and the normal region in the image may not be significantly different is not considered, resulting in poor accuracy of the segmentation results obtained by directly segmenting the filtered lower limb CT images. Summary of the Invention

[0004] In order to solve the technical problem of inaccurate segmentation results of lower limb CT images in the prior art, the purpose of the present invention is to provide an intelligent image segmentation system for detecting lower limb trauma fractures, and the specific technical solutions adopted are as follows: The present invention provides an intelligent image segmentation system for detecting lower limb trauma fractures, and the system includes: An image acquisition module for acquiring lower limb CT images; A pixel point abnormality acquisition module for obtaining the change degree of each pixel point according to the difference in the change rules of the pixel values of each pixel point in the lower limb CT image in the row and column directions; and obtaining the abnormality of each pixel point according to the fluctuation influence of the change degree of each pixel point in the row where each pixel point is located; A local abnormality analysis module for obtaining the adjacent regions of the preset neighborhood region corresponding to each pixel point within the local range of the preset neighborhood region corresponding to each pixel point; and obtaining the trauma possibility of each pixel point according to the abnormality difference between the preset neighborhood region corresponding to each pixel point and each adjacent region, and the symmetry of the adjacent regions; An image segmentation module for obtaining an enhanced lower limb CT image according to the trauma possibility of each pixel point and segmenting the enhanced lower limb CT image.

[0005] Further, the method for obtaining the change degree includes: In each row of pixel points of the lower limb CT image, obtain the position index value of each pixel point in the order from left to right; combine the position index values of each pixel point to obtain the corresponding pixel sequence for each row in the lower limb CT image; the pixel points in the pixel sequence are arranged in ascending order according to the pixel values, and the pixel points with the same pixel value are arranged in ascending order according to the position index values; Obtain the pixel index value of each pixel point according to the arrangement order of pixel points in the pixel sequence; Successively take each pixel point in the lower limb CT image as a reference pixel point, and take the row where the reference pixel point is located as the target row; in each row adjacent to the target row, obtain the pixel index value of the pixel point with the same position index value as the reference pixel point as the adjacent index value of the reference pixel point; calculate the difference between the pixel index value of the reference pixel point and the adjacent index value to obtain the adjacent difference of the reference pixel point. When the number of adjacent differences of the reference pixel point is 1, take the preset change degree as the change degree of the reference pixel point; when the number of adjacent differences of the reference pixel point is greater than 1, take the difference between the adjacent differences of the reference pixel point as the change degree of the reference pixel point.

[0006] Further, the method for obtaining the abnormality includes: Successively take each row in the lower limb CT image as a reference row. For any pixel point in the reference row, calculate the variance of the change degrees of all other pixel points in the reference row except this pixel point to obtain the change difference degree of this pixel point. Obtain the mean value of the change difference degrees corresponding to all pixel points in the reference row, and take the difference between the change difference degree of each pixel point in the reference row and the mean value of the change difference degrees as the change difference coefficient of each pixel point. Calculate the product of the change difference coefficient of each pixel point and the change degree to obtain the abnormality of each pixel point.

[0007] Further, the method for obtaining the adjacent region includes: Successively take each pixel point in the lower limb CT image as the central pixel point, and take the preset neighborhood area corresponding to the center point as the central area; Obtain the nearest adjacent regions corresponding to the central area in different preset directions of the central area. The shape and size of the adjacent regions are the same as those of the central area, and the pixel points in the adjacent regions do not overlap with the pixel points in the central area.

[0008] Further, the method for obtaining the trauma possibility includes: For any central area, in the central area and all corresponding adjacent areas, sort the abnormalities of the pixel points according to the preset arrangement order of the pixel point positions to obtain the abnormality sequences corresponding to the central area and each adjacent area; Take the two adjacent areas in each preset symmetry direction of the central area as a pair of symmetric areas; Successively take each adjacent area in the central area as the detection area, and obtain the abnormal difference of the central area corresponding to the detection area according to the difference between the abnormality sequences corresponding to the central area and the detection area. In each pair of symmetric regions corresponding to the central region, the difference between the abnormal differences of two adjacent regions is used as the regional difference of the central region; calculate the cumulative value of all regional differences corresponding to the central region to obtain the trauma possibility of the central pixel point corresponding to the central region.

[0009] Further, the specific expression of the abnormal difference is: ; In the formula, represents the abnormal difference between the th and th adjacent regions corresponding to the central region ; represents the th abnormality in the abnormality sequence of the central region ; represents the th abnormality in the abnormality sequence of the th adjacent region ;

[0010] Further, obtaining the enhanced lower limb CT image according to the trauma possibility of each pixel point includes: Taking the trauma possibility after negative correlation mapping and normalization as the pixel adjustment coefficient of each pixel point; Calculating the product of the maximum gray level of the pixel point and the pixel adjustment coefficient of each pixel point to obtain the new pixel value of each pixel point, thereby obtaining the enhanced lower limb CT image.

[0011] Further, the method for obtaining the nearest adjacent region includes: In each preset direction, calculate the distance between the adjacent central point of the adjacent region and the central pixel point of the central region, and take the adjacent region corresponding to the adjacent central point with the smallest distance as the nearest adjacent region.

[0012] Further, the preset neighborhood region is an eight-neighborhood region.

[0013] Further, the Otsu threshold segmentation algorithm is used to segment the enhanced lower limb CT image.

[0014] The present invention has the following beneficial effects: The present invention analyzes the variation law of pixel points in the row and column of lower limb CT images to obtain the variation degree of each pixel point. The variation degrees of pixel points in each row in the bone region are similar, but fractures will disrupt the similarity of the variation degrees. Therefore, according to the influence of the variation degree of each pixel point on the fluctuation in the row where the pixel point is located, the abnormality of each pixel point is obtained, and the abnormality can reflect the possibility that the pixel point is in a traumatic fracture region. However, when there is an overlap in the lower limb leg bone due to a bone crack, the abnormality of the pixel point instead approaches that of a normal pixel point. Therefore, it is necessary to further analyze the regional abnormality of the local area range of each pixel point, that is, to obtain the adjacent regions corresponding to the preset neighborhood region of each pixel point. According to the abnormality difference between the preset neighborhood region corresponding to each pixel point and each adjacent region, as well as the symmetry of the adjacent regions, the trauma possibility of each pixel point is obtained, and the possibility that the pixel point is in the traumatic fracture region is analyzed more comprehensively, making the enhancement accuracy of the pixel value of the pixel point higher, obtaining a greater contrast between the pixel points in the enlarged trauma region and the normal region in the lower limb enhanced CT image, making the reliability of the segmentation result obtained by finally segmenting the lower limb enhanced CT image stronger, and obtaining a more accurate and clear image segmentation result. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 It is a structural diagram of an intelligent image segmentation system for detecting traumatic fractures in the lower limb provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in combination with the drawings and preferred embodiments, details the specific implementation manners, structures, features, and effects of an intelligent image segmentation system for detecting traumatic fractures in the lower limb proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the 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 those of ordinary skill in the technical field to which the present invention belongs.

[0019] The following specifically describes the specific solution of an intelligent image segmentation system for detecting lower limb trauma fractures provided by the present invention in combination with the accompanying drawings.

[0020] Please refer to Figure 1 , which shows the structure diagram of an intelligent image segmentation system for detecting lower limb trauma fractures provided by an embodiment of the present invention. The system includes: an image acquisition module 101, a pixel abnormality acquisition module 201, a local abnormality analysis module 301, and an image segmentation module 401.

[0021] The image acquisition module 101 is used to acquire lower limb CT images.

[0022] In an embodiment of the present invention, an initial lower limb CT image can be collected using a CT scanner. It should be noted that the initial lower limb CT image is a grayscale image after generation and does not need to be grayscale processed again. Moreover, the lower limb is usually the bone of the leg bone, which is arranged vertically in the CT image. Further, the initial lower limb CT image is denoised using Gaussian filtering to obtain the lower limb CT image. It should be noted that the acquisition of the CT image and Gaussian filtering denoising are well-known technical means to those skilled in the art and will not be elaborated here.

[0023] The pixel abnormality acquisition module 201 is used to obtain the change degree of each pixel point according to the difference in the change rules of the pixel values of each pixel point in the row and column of the lower limb CT image; and obtain the abnormality of each pixel point according to the fluctuation influence of the change degree of each pixel point in the row where each pixel point is located.

[0024] Since in the lower limb CT image, the leg bone is arranged vertically, and the pixel points in the bone area have continuous and similar pixel value change characteristics. For the pixel points in the fractured area affected by trauma, since the fracture will destroy the similarity of the pixel value changes corresponding to the pixel points, first, the change degree of each pixel point is obtained according to the difference in the change rules of the pixel values of each pixel point in the row and column of the lower limb CT image.

[0025] Preferably, in each row of pixel points of the lower limb CT image, the position index value of each pixel point is obtained in the order from left to right, and the corresponding pixel sequence of each row in the lower limb CT image is obtained in combination with the position index value of each pixel point. Among them, the pixel values of the pixel points in the pixel sequence are arranged from small to large, and the pixel points with the same pixel value are arranged from small to large according to the position index value. By combining the pixel value sequence of the pixel point with the arrangement position information of the pixel point in each row, the pixel sequence is obtained. The pixel sequence of each row mainly reflects the distribution arrangement rule of the pixel values of the pixel points in each row. After combining with the position index value of the pixel point, the distribution change of the pixel values when the column numbers of the pixel points are the same can be analyzed.

[0026] Obtain the pixel index value of each pixel point according to the arrangement order of pixel points in the pixel sequence. The pixel index value reflects the pixel value distribution information of each pixel point. Successively take each pixel point in the lower limb CT image as a reference pixel point, and take the row where the reference pixel point is located as the target row. In each row adjacent to the target row, obtain the pixel index value of the pixel point with the same position index value as the reference pixel point, as the adjacent index value of the reference pixel point. The adjacent index value is the pixel value distribution information of the adjacent pixel points in the same column as the reference pixel point.

[0027] Preferably, calculate the difference between the pixel index value of the reference pixel point and the adjacent index value to obtain the adjacent difference of the reference pixel point. The adjacent difference is the difference in the pixel value distribution information between the reference pixel point and an adjacent pixel point with the same number of columns.

[0028] When the number of adjacent differences of the reference pixel point is 1, it indicates that the position of the reference pixel point is in the first row and the last row of the lower limb CT image, and there is only one row adjacent to the target row. Since in actual CT shooting, the trauma position to be analyzed will not be in the first row and the last row of the image, the preset change degree is used as the change degree of the reference pixel point. In the embodiment of the present invention, the preset change degree is 0, and the specific value can be adjusted by the implementer himself.

[0029] When the number of adjacent differences of the reference pixel point is greater than 1, that is, the reference pixel point is located in other rows except the first row and the last row, and there are two rows adjacent to the target row. Take the difference between the adjacent differences of the reference pixel point as the change degree of the reference pixel point. In the embodiment of the present invention, the specific expression of the change degree is: ; In the formula, represents the change degree of the th pixel point, represents the pixel index value of the th pixel point in the th row, represents the adjacent index value of the th pixel point corresponding to the th row, represents the adjacent index value of the th pixel point corresponding to the th row, represents the absolute value extraction function.

[0030] Among them, and both represent the adjacent differences of the th pixel point. In the embodiment of the present invention, the calculation of the difference is the absolute value of the difference. The change degree reflects the change degree of the pixel value distribution arrangement of each pixel point and the adjacent pixel points in the same column in the adjacent rows.

[0031] Further considering that the bones of the leg bones are vertically distributed, the degree of change of the pixel points in each row is similar. However, when the pixel points in the row are the pixel points corresponding to the trauma fracture area, the change situation of the pixel points in the trauma fracture area is inconsistent with the regularity of the normal pixel points, and the difference in the degree of change of the pixel points in the trauma fracture area from the overall regularity of the row where they are located is relatively large.

[0032] In the embodiment of the present invention, each row in the lower limb CT image is sequentially used as a reference row. For any pixel point in the reference row, the variance of the degree of change of all other pixel points in the reference row except this pixel point is calculated to obtain the change difference degree of this pixel point. The contribution degree of this pixel point to the variance of the degree of change of the entire row is reflected by the change difference degree.

[0033] The mean value of the corresponding change difference degrees of all pixel points in the reference row is obtained. The average contribution degree of each pixel point to the variance is reflected by the mean value. The difference between the change difference degree of each pixel point in the reference row and the mean value of the change difference degrees is used as the change difference coefficient of each pixel point. When the change difference coefficient is larger, it indicates that the influence degree of the pixel point on the variance is greater, and the abnormality degree of this pixel point is more obvious.

[0034] The product of the change difference coefficient and the degree of change of each pixel point is calculated to obtain the abnormality of each pixel point. In the embodiment of the present invention, the specific expression of the abnormality is: ; In the formula, represents the abnormality of the th pixel point, represents the degree of change of the th pixel point, represents the mean value of the change difference degrees of the row where the th pixel point is located, represents the change difference degree of the th pixel point, represents the absolute value extraction function.

[0035] Among them, represents the change difference coefficient of the th pixel point. The abnormality reflects the possibility that the pixel point is a pixel point in the trauma fracture area. When the pixel point belongs to the pixel point in the normal area, the difference between the degree of change of the pixel point and the overall regularity of the row where it is located is not large, and the overall regularity of the pixel points in each row is relatively consistent. Therefore, the abnormality of the pixel point will be smaller. When the pixel point belongs to the pixel point in the trauma fracture area, the difference between the degree of change of the pixel point and the overall regularity of the row where it is located is relatively large, and the abnormality of the pixel point will also be larger.

[0036] The local anomaly analysis module 301 is used to obtain the adjacent regions of the preset neighborhood area corresponding to each pixel point within the local range of the preset neighborhood area corresponding to each pixel point; according to the anomaly difference between the preset neighborhood area corresponding to each pixel point and each adjacent region, and the symmetry of the adjacent regions, obtain the trauma possibility of each pixel point.

[0037] By analyzing the regular characteristics of the rows and columns of pixel points in the lower limb CT image, a preliminary judgment of the anomaly of the pixel points is obtained to estimate the possibility of the pixel points in the trauma fracture area. However, considering that there are still some pixel points in the trauma fracture area with unclear anomaly. For example, when a bone crack occurs in the leg bone, sometimes a fracture bone superimposed area will be generated. When there is a superposition of bone fragments, the regularity of the local row and column change degree is not significantly damaged, and the anomaly tends to be a normal pixel point. Therefore, simply relying on regularity cannot fully determine that the pixel point is a pixel point in the trauma bone area, and it is necessary to consider the influence of the overall difference degree of each pixel point within the local range.

[0038] Therefore, a preset neighborhood area is set for each pixel point for analysis. In the embodiment of the present invention, the preset neighborhood area is an eight-neighborhood area centered on each pixel point. Each pixel point in the lower limb CT image is sequentially used as the central pixel point, and the preset neighborhood area corresponding to the central point is used as the central area. The central area contains 9 pixel points. The nearest adjacent areas corresponding to the central area are obtained in different preset directions of the central area. The shape and size of the adjacent areas are the same as those of the central area, and the pixel points in the adjacent areas do not overlap with the pixel points in the central area. In the embodiment of the present invention, the preset direction is the 8-chain code direction, that is, an adjacent area with the same shape and size is obtained in each direction of the 8-chain code for the central area. Each adjacent area contains 9 pixel points. To ensure that the adjacent area is the nearest adjacent area, in each preset direction, the distance between the adjacent central point of the adjacent area and the central pixel point of the central area is calculated, and the adjacent area corresponding to the adjacent central point with the smallest distance is used as the nearest adjacent area. It should be noted that the calculation of the distance can use the Euclidean distance, which is a well-known technical means to those skilled in the art and will not be elaborated here.

[0039] Through the overall difference analysis between regions in the local range of the central region, more accurate judgment of pixel points can be carried out. Preferably, for any central region, among the central region and all corresponding adjacent regions, the abnormality of pixel points is sorted according to the preset arrangement order according to the pixel point positions, and the abnormality sequence corresponding to the central region and each adjacent region is obtained. In the embodiment of the present invention, the preset arrangement order is set as follows: taking the pixel point at the upper left corner position of each region as the starting point, sorting in the order from left to right in each row, and taking the pixel point at the lower right corner position of the region as the end point. According to this arrangement order, the abnormality of pixel points is sorted in sequence to obtain the abnormality sequence. The preset arrangement order can be adjusted by the implementer himself, as long as it is ensured that the abnormality in each region is arranged in the same traversal order of pixel point positions, so as to ensure the difference between abnormalities during calculation. The abnormality is the abnormality corresponding to the pixel point positions in two regions.

[0040] Taking the two adjacent regions in each preset symmetric direction of the central region as a symmetric region pair. In the embodiment of the present invention, the preset symmetric directions are the horizontal direction, the vertical direction, the 45-degree diagonal direction, and the 135-degree diagonal direction. When an adjacent region can be obtained in each direction of the 8-chain code of the central region, there are 4 symmetric region pairs, and two adjacent regions can be obtained in each symmetric direction to form a symmetric region pair.

[0041] Successively taking each adjacent region in the central region as the detection region, according to the difference between the abnormality sequences corresponding to the central region and the detection region, the abnormality difference corresponding to the central region for the detection region is obtained. In the embodiment of the present invention, the expression of the abnormality difference is: ; In the formula, represents the abnormality difference corresponding to the th adjacent region of the central region , represents the th abnormality in the abnormality sequence of the central region , represents the th abnormality in the abnormality sequence of the th adjacent region, represents the total number of abnormalities in the abnormality sequence, represents the absolute value extraction function.

[0042] The deviation change of the abnormality between the central region and each adjacent region is reflected by the abnormality difference. Further, in each pair of symmetric regions corresponding to the central region, the difference between the abnormality differences of the two adjacent regions is used as the regional difference of the central region. The regional difference reflects the degree of deviation of the difference in each symmetric direction of the central region. The greater the regional difference, the more likely it is that there is a traumatic fracture region in the local range. Calculate the cumulative value of all regional differences corresponding to the central region to obtain the trauma possibility of the central pixel point corresponding to the central region. In the embodiment of the present invention, the specific expression of the trauma possibility: ; In the formula, represents the trauma possibility of the th pixel point, represents the abnormality difference between the central region in the th pair of symmetric regions and the first adjacent region, represents the abnormality difference between the central region in the th pair of symmetric regions and the second adjacent region, represents the total number of pairs of symmetric regions, represents the absolute value extraction function.

[0043] Among them, represents the regional difference of the central region corresponding to the th pair of symmetric regions, represents the central region with the th pixel point as the central pixel point. The greater the regional difference, the greater the degree of fluctuation of the abnormality in the local range corresponding to the central region, and the greater the overall deviation degree of the abnormality, indicating that the corresponding central pixel point is more likely to be a pixel point in the traumatic fracture region.

[0044] The image segmentation module 401 is used to obtain the enhanced CT image of the lower limb according to the trauma possibility of each pixel point and segment the enhanced CT image of the lower limb.

[0045] The trauma possibility of pixel points can be further used to enhance the lower limb CT image, increasing the pixel value difference between the pixel points in the trauma fracture area and the normal pixel points, so that the trauma fracture area can be more accurately segmented during subsequent segmentation. Since the pixel points in the trauma fracture area usually have lower pixel values, in the embodiments of the present invention, the trauma possibility after negative correlation mapping and normalization processing is used as the pixel adjustment coefficient for each pixel point. The greater the trauma possibility of the pixel point, the more likely the pixel point is a pixel point in the trauma fracture area, and the smaller the adjustment coefficient of the pixel point. Further, the product of the maximum gray level of the pixel point and the pixel adjustment coefficient of each pixel point is calculated to obtain the new pixel value of each pixel point. The smaller the adjustment coefficient, the smaller the new pixel value of the pixel point, and the greater the difference between the pixel points in the trauma fracture area and the normal pixel points, thus obtaining the enhanced lower limb CT image.

[0046] In other embodiments of the present invention, it is also possible to screen out the pixel points that may be in the trauma fracture area by setting a trauma threshold, and increase the pixel value contrast of the pixel points in the trauma fracture area by reducing the pixel values of the pixel points in the trauma fracture area, which will not be elaborated here.

[0047] The Otsu threshold segmentation algorithm is used to segment the enhanced lower limb CT image. In the embodiments of the present invention, the segmentation threshold can be obtained through the Otsu threshold segmentation algorithm. The pixel value of the pixel points smaller than the segmentation threshold is set to 0, and the pixel value of the pixel points greater than the segmentation threshold is recorded as 255, obtaining a binary image. The area with a pixel value of 0 in the binary image reflects the trauma fracture area. It should be noted that the Otsu threshold segmentation algorithm is a well-known technical means for those skilled in the art and will not be elaborated here.

[0048] Thus, by increasing the contrast between the pixel points in the trauma fracture area and the normal pixel points, more accurate image segmentation is performed.

[0049] In summary, the present invention analyzes the variation law of pixel points in the row and column of lower limb CT images to obtain the variation degree of each pixel point. The variation degrees of pixel points in each row in the bone region are similar, but fractures will cause the similarity of the variation degrees to be destroyed. Therefore, according to the fluctuation influence of the variation degree of each pixel point in the row where each pixel point is located, the abnormality of each pixel point is obtained. The abnormality can reflect the possibility that the pixel point is in the traumatic fracture region. However, when there is a bone crack in the lower limb leg bone, there will be an overlapping situation. At this time, the abnormality of the pixel point instead approaches that of a normal pixel point. Therefore, it is necessary to further analyze the regional abnormality of the local area range of each pixel point, that is, obtain the adjacent regions corresponding to the preset neighborhood region of each pixel point. According to the abnormality difference between the preset neighborhood region corresponding to each pixel point and each adjacent region, and the symmetry of the adjacent regions, the trauma possibility of each pixel point is obtained, and the possibility that the pixel point is in the traumatic fracture region is analyzed more comprehensively, so that the accuracy of enhancing the pixel value of the pixel point is higher, the contrast between the pixel points in the enlarged trauma region and the normal region of the lower limb enhanced CT image is obtained, and the lower limb enhanced CT image is segmented to obtain a more accurate and clear image segmentation result.

[0050] It should be noted that the above sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0051] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other, and the focus of each embodiment is to illustrate the differences from other embodiments.

Claims

1. An intelligent image segmentation system for detecting lower limb traumatic fractures, characterized in that, The system includes: an image acquisition module for acquiring lower limb CT images; a pixel abnormality acquisition module for obtaining the change degree of each pixel according to the difference in the change rules of the pixel values of each pixel in the lower limb CT image in the row and column directions; and obtaining the abnormality of each pixel according to the fluctuation influence of the change degree of each pixel in the row where each pixel is located; a local abnormality analysis module for obtaining the adjacent region corresponding to the preset neighborhood region of each pixel within the local range corresponding to the preset neighborhood region of each pixel; and obtaining the trauma possibility of each pixel according to the abnormality difference between the preset neighborhood region corresponding to each pixel and each adjacent region, and the symmetry of the adjacent region; an image segmentation module for obtaining an enhanced lower limb CT image according to the trauma possibility of each pixel and segmenting the enhanced lower limb CT image.

2. The intelligent image segmentation system for detecting lower limb trauma fractures according to claim 1, wherein The method for obtaining the change degree includes: In each row of pixel points in the lower limb CT image, obtaining the position index value of each pixel point in the order from left to right; combining the position index value of each pixel point to obtain the pixel sequence corresponding to each row in the lower limb CT image; the pixel points in the pixel sequence are arranged in ascending order of pixel value, and the pixel points with the same pixel value are arranged in ascending order of position index value; obtaining the pixel index value of each pixel point according to the arrangement order of the pixel points in the pixel sequence; successively taking each pixel point in the lower limb CT image as a reference pixel point and the row where the reference pixel point is located as the target row; in each row adjacent to the target row, obtaining the pixel index value of the pixel point with the same position index value as the reference pixel point as the adjacent index value of the reference pixel point; calculating the difference between the pixel index value of the reference pixel point and the adjacent index value to obtain the adjacent difference of the reference pixel point; when the number of adjacent differences of the reference pixel point is 1, taking the preset change degree as the change degree of the reference pixel point; when the number of adjacent differences of the reference pixel point is greater than 1, taking the difference between the adjacent differences of the reference pixel point as the change degree of the reference pixel point.

3. The intelligent image segmentation system for detecting lower limb traumatic fractures according to claim 1, wherein The method for obtaining the abnormality includes: successively taking each row in the lower limb CT image as a reference row, and for any pixel point in the reference row, calculating the variance of the change degrees of all other pixel points in the reference row except this pixel point to obtain the change difference degree of this pixel point; obtaining the mean value of the change difference degrees corresponding to all pixel points in the reference row, and taking the difference between the change difference degree of each pixel point in the reference row and the mean value of the change difference degrees as the change difference coefficient of each pixel point; calculating the product of the change difference coefficient of each pixel point and the change degree to obtain the abnormality of each pixel point.

4. The intelligent image segmentation system for detecting lower limb traumatic fractures according to claim 1, wherein, The method for obtaining the adjacent region includes: successively taking each pixel point in the lower limb CT image as a central pixel point and the preset neighborhood region corresponding to the central point as the central region; obtaining the nearest adjacent region corresponding to the central region in different preset directions of the central region, the shape and size of the adjacent region are the same as those of the central region, and the pixel points in the adjacent region do not overlap with the pixel points in the central region.

5. The intelligent image segmentation system for detecting lower limb traumatic fractures according to claim 4, characterized in that, The method for obtaining the trauma possibility includes: For any central region, among the central region and all its corresponding adjacent regions, sort the abnormality of pixel points according to the preset arrangement order based on the pixel point positions to obtain the abnormality sequences corresponding to the central region and each adjacent region; Take two adjacent regions in each preset symmetry direction of the central region as a symmetry region pair; Successively take each adjacent region in the central region as a detection region, and obtain the abnormality difference corresponding to the detection region of the central region according to the difference between the abnormality sequences corresponding to the central region and the detection region; In each symmetry region pair corresponding to the central region, take the difference between the abnormality differences of the two adjacent regions as the region difference of the central region; calculate the cumulative value of all region differences corresponding to the central region to obtain the trauma possibility of the central pixel point corresponding to the central region.

6. The intelligent image segmentation system for detecting lower limb traumatic fractures according to claim 5, wherein, The specific expression of the abnormality difference is: ； Wherein, represents the central region corresponding to the abnormal difference between the nth adjacent regions, the nth abnormality in the abnormality sequence of the central region represents the nth abnormality in the abnormality sequence of the nth adjacent region, represents the total number of abnormalities in the abnormality sequence, represents the absolute value extraction function.

7. The intelligent image segmentation system for detecting lower limb traumatic fractures according to claim 1, wherein Obtaining the enhanced CT image of the lower limb according to the trauma possibility of each pixel point includes: Take the trauma possibility after negative correlation mapping and normalization processing as the pixel adjustment coefficient of each pixel point; Calculate the product of the maximum gray level of the pixel point and the pixel adjustment coefficient of each pixel point to obtain the new pixel value of each pixel point, and obtain the enhanced CT image of the lower limb.

8. The intelligent image segmentation system for detecting lower limb trauma fractures according to claim 4, wherein The method for obtaining the nearest adjacent region includes: In each preset direction, calculate the distance between the adjacent central point of the adjacent region and the central pixel point of the central region, and take the adjacent region corresponding to the adjacent central point with the smallest distance as the nearest adjacent region.

9. The intelligent image segmentation system for detecting lower limb trauma fractures according to claim 1, characterized in that The preset neighborhood region is an eight-neighborhood region.

10. The intelligent image segmentation system for detecting lower limb traumatic fractures according to claim 1, characterized in that, Use the Otsu threshold segmentation algorithm to segment the enhanced CT image of the lower limb.

Images