Intelligent segmentation system of knee joint images based on local threshold

Abstract

The invention relates to the field of image processing technology, and in particular to a local threshold-based intelligent segmentation system for knee joint images, comprising: a knee joint grayscale image acquisition module, a retention rate acquisition module for each pixel, a retention degree acquisition module for each image block, an evaluation value acquisition module for each adjusted segmentation threshold, and an image segmentation module, which acquires a knee joint grayscale image, performs block processing on the knee joint grayscale image to obtain a plurality of image blocks, calculates the retention rate of each pixel, obtains the retention degree of each image block according to the retention rate of each pixel, obtains the evaluation value of each adjusted segmentation threshold according to the retention degree of each image block, and then obtains an optimal segmentation threshold, performs segmentation processing on the knee joint grayscale image according to the optimal segmentation threshold to obtain a knee joint area, thereby obtaining an accurate segmentation threshold and improving the accuracy of knee joint image segmentation.

Description

Technical Field

[0001] The invention relates to the technical field of image processing, and in particular to a knee joint image intelligent segmentation system based on local threshold value. Background Art

[0002] At present, the knee joint images are generally obtained through magnetic resonance imaging technology, which can efficiently collect images without affecting the health of the human body.

[0003] Due to the imperfection of technology and the differences in body structure, the collected knee joint images will inevitably have some uneven brightness fields, which will cause the same tissue to have different grayscale ranges in the image, and it is impossible to use the same threshold to segment the complete knee joint area. Therefore, the knee joint images segmented by the traditional threshold segmentation method are not complete, which can easily cause trouble for doctors in diagnosis.

[0004] In image processing, for images with uneven lighting, it is difficult to segment the complete image area using the global threshold due to the problem of non-uniform grayscale. At the same time, for the local area threshold segmentation method, since the pixels in the local area are small, the segmentation threshold obtained by statistical features is not accurate enough, making it difficult to accurately segment the image area. Summary of the invention

[0005] The present invention provides a knee joint image intelligent segmentation system based on local threshold value to solve the existing problem: obtaining an accurate segmentation threshold value for each small area, thereby segmenting a complete knee joint area.

[0006] The local threshold-based intelligent segmentation system for knee joint images of the present invention adopts the following technical solutions:

[0007] An embodiment of the present invention provides a knee joint image intelligent segmentation system based on local threshold, the system comprising the following modules:

[0008] A knee joint grayscale image acquisition module, used for acquiring a knee joint grayscale image;

[0009] A retention rate acquisition module for each pixel is used to process the knee joint grayscale image into blocks to obtain multiple image blocks, set multiple adjustment segmentation thresholds for each image block, use each adjustment segmentation threshold to perform segmentation processing on each image block to obtain a binary image block of each image block under each adjustment segmentation threshold, and obtain the retention rate of each pixel according to the change of the grayscale value of each pixel in two binary image blocks under two adjacent adjustment segmentation thresholds;

[0010] A retention degree acquisition module for each image block, used to obtain the change weight of each pixel according to the number of gray value change pixels in two adjacent binary image blocks under the adjusted segmentation threshold, and obtain the retention degree of each image block according to the change weight of each pixel and the retention rate of each pixel;

[0011] An evaluation value acquisition module for each adjusted segmentation threshold, used to obtain an evaluation value for each adjusted segmentation threshold according to the degree of retention of each image block and the consistency of the mean grayscale value in the binary image block of each image block under each adjusted segmentation threshold;

[0012] The image segmentation module is used to obtain the optimal segmentation threshold according to the evaluation value of each adjusted segmentation threshold, and to segment the knee joint grayscale image according to the optimal segmentation threshold to obtain the knee joint area.

[0013] Preferably, the grayscale image of the knee joint is processed into blocks to obtain a plurality of image blocks, and the specific method includes:

[0014] The knee joint grayscale image is evenly divided into multiple image blocks.

[0015] Preferably, the setting of multiple adjustment segmentation thresholds for each image block includes the following specific methods:

[0016] For any image block, the Otsu threshold algorithm is used to obtain the baseline segmentation threshold of the image block. , set the variance adjustment variable , Gaussian variance , let k take all integers between [0,20] to obtain multiple Gaussian variances, b represents the preset adjustment step size, and each adjusted segmentation threshold is obtained based on each Gaussian variance: ,in Indicates adjusting the segmentation threshold. Represents an exponential function with a natural constant as its base.

[0017] Preferably, the method of obtaining the retention rate of each pixel according to the change of the gray value of each pixel in two adjacent binary image blocks under the adjusted segmentation thresholds includes the following specific methods:

[0018] Arrange the adjusted segmentation thresholds in ascending order according to the size of the Gaussian variance to obtain an adjusted segmentation threshold sequence;

[0019] For any image block, any adjusted segmentation threshold is recorded as the first target segmentation threshold, a pixel at any position is recorded as a target pixel, the grayscale value at the target pixel in the binary image block of the first target segmentation threshold is recorded as the first grayscale value of the target pixel, and the grayscale value at the target pixel in the binary image block of the previous adjusted segmentation threshold is recorded as the second grayscale value of the target pixel. When the first grayscale value is equal to 1 and the second grayscale value is equal to 0, the Gaussian variance corresponding to the first target segmentation threshold is used as the retention rate of the target pixel to obtain the retention rate of each pixel in the image block.

[0020] Preferably, the specific method of obtaining the change weight of each pixel according to the number of pixels with changed gray values ​​in two adjacent binary image blocks under the adjusted segmentation thresholds includes:

[0021] For any image block, any adjusted segmentation threshold is recorded as the second target segmentation threshold, the number of pixels with a gray value of 1 in the binary image block under the second target segmentation threshold is recorded as the first number of the second target segmentation threshold, the number of pixels with a gray value of 1 in the binary image block of the previous adjusted segmentation threshold is recorded as the second number of the second target segmentation threshold, and the difference between the first number and the second number is recorded as the variation difference of the second target segmentation threshold; the variation difference of each adjusted segmentation threshold is obtained, the Gaussian variance with the same retention rate as each pixel is recorded as the base Gaussian variance, the adjusted segmentation threshold corresponding to the base Gaussian variance is used as the base adjusted segmentation threshold of each pixel, and the variation difference of the base adjusted segmentation threshold of each pixel is used as the variation weight of each pixel;

[0022] Get the change weights for each pixel in each image patch.

[0023] Preferably, the method of obtaining the retention degree of each image block according to the change weight of each pixel and the retention rate of each pixel includes the following specific methods:

[0024]

[0025] in, represents the retention rate of the i-th pixel in the z-th image block; represents the change weight of the i-th pixel in the z-th image block, n represents the number of pixels in each image block, Indicates the degree of preservation of the z-th image block.

[0026] Preferably, the evaluation value of each adjusted segmentation threshold is obtained according to the consistency between the retention degree of each image block and the mean gray value of each image block in the binary image block under each adjusted segmentation threshold, including the specific method of:

[0027] The mean gray value of all pixels in the binary image block under each adjusted segmentation threshold is recorded as the mean gray value of the image block under each adjusted segmentation threshold;

[0028] The calculation method for obtaining the evaluation value of each adjusted segmentation threshold according to the mean gray value of each image block at each adjusted segmentation threshold and the retention degree of each image block is:

[0029]

[0030] in, Indicates the degree of preservation of the z-th image block, represents the mean gray value of the z-th image block under each adjusted segmentation threshold, represents the number of image blocks, P represents the evaluation value of each adjusted segmentation threshold, and exp() represents an exponential function with a natural constant as the base.

[0031] Preferably, the method of obtaining the optimal segmentation threshold according to the evaluation value of each adjusted segmentation threshold includes:

[0032] The adjusted segmentation threshold with the largest evaluation value is taken as the optimal segmentation threshold.

[0033] Preferably, the method of performing segmentation processing on each image block using each adjusted segmentation threshold to obtain a binary image block of each image block under each adjusted segmentation threshold includes the following specific methods:

[0034] The grayscale values ​​of pixels in each image block whose grayscale values ​​are greater than each adjusted segmentation threshold are set to 1, and the grayscale values ​​of pixels whose grayscale values ​​are less than or equal to each adjusted segmentation threshold are set to 0, thereby obtaining a binary image block of each image block under each adjusted segmentation threshold.

[0035] Preferably, the step of segmenting the knee joint grayscale image according to the optimal segmentation threshold to obtain the knee joint region comprises the following specific methods:

[0036] The grayscale values ​​of pixels in the knee joint grayscale image that are greater than the optimal segmentation threshold are set to 1, and the grayscale values ​​of pixels in the knee joint grayscale image that are less than or equal to the optimal segmentation threshold are set to 0 to obtain an optimal binary image, wherein the pixels in the optimal binary image whose grayscale values ​​are equal to 1 are knee joint pixels;

[0037] The connected domain formed by knee joint pixels is regarded as the knee joint area.

[0038] The technical solution of the present invention has the following beneficial effects: since there are differences in the grayscale values ​​of knee joint pixels in various regions, it is difficult to directly obtain the segmentation threshold of the entire image, and the analysis difficulty of each image block is relatively low, so the knee joint grayscale image is divided into blocks to obtain multiple image blocks, and since the grayscale values ​​of knee joint pixels are relatively large, when different thresholds are used for segmentation, the knee joint pixels will be larger than the segmentation threshold in most cases, and thus multiple different adjustment segmentation thresholds are set, and each image block is segmented using multiple different adjustment segmentation thresholds to obtain a binary image block under each adjustment segmentation threshold, and multiple binary images are analyzed to obtain the retention rate of each pixel according to the change of the grayscale value of each pixel in the binary image block, and since the number of pixels with grayscale value changes in the binary image block caused by the change of each adjustment segmentation threshold is different, the corresponding degree of influence of the adjustment segmentation threshold is different, and thus the change weight of each pixel is obtained according to the number of pixels with grayscale value changes in the binary image block caused by the change of each adjustment segmentation threshold; and the retention degree of each image block is obtained according to the change weight and the retention rate. The retention degree can reflect the probability that each image block is the knee joint area, where the grayscale value at the knee joint area is larger. Therefore, the larger the grayscale value mean of the image block, the greater the probability that it is the knee joint area. Therefore, the evaluation value of each adjusted segmentation threshold is obtained according to the consistency between the grayscale value mean of each image block under each adjustment threshold and the retention degree of the image block. The optimal segmentation threshold is obtained according to the evaluation value, and the accurate knee joint pixel area is segmented out using the optimal segmentation threshold. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 This is a structural block diagram of the intelligent segmentation system of knee joint images based on local threshold value of the present invention. DETAILED DESCRIPTION

[0041] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of the intelligent segmentation system of knee joint images based on local thresholds proposed by the present invention are described in detail as follows in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0042] 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.

[0043] The specific scheme of the intelligent segmentation system of knee joint images based on local threshold provided by the present invention is described in detail below with reference to the accompanying drawings.

[0044] See also Figure 1 , which shows a structural block diagram of a local threshold-based intelligent segmentation system for knee joint images provided by an embodiment of the present invention, the system includes the following modules:

[0045] The knee joint grayscale image acquisition module obtains the patient's knee joint grayscale image.

[0046] It should be noted that due to the influence of factors such as uneven light, different grayscale values ​​exist in the knee joints in different areas. Therefore, it is difficult to accurately segment the complete knee joint area by using the same threshold to segment the entire image. As for the block threshold segmentation method, since there are fewer pixels in each block, it is impossible to obtain accurate segmentation thresholds for each block using statistical features, and thus the knee joint area cannot be segmented. For this reason, this embodiment proposes a knee joint image intelligent segmentation system based on local thresholds.

[0047] In order to realize the local threshold-based intelligent segmentation system of knee joint images proposed in this embodiment, it is first necessary to collect a grayscale image of the knee joint. The specific process is: obtain the patient's knee joint image in the hospital's image resource library, and grayscale the knee joint image to obtain a grayscale image of the knee joint.

[0048] So far, the grayscale image of the knee joint is obtained through the above method.

[0049] The retention rate acquisition module for each pixel divides the grayscale image of the knee joint into several image blocks, uses different adjusted segmentation thresholds to segment each image block to obtain several binary image blocks of each image block, and obtains the retention rate of each pixel according to the grayscale changes of the several binary image blocks of each segment.

[0050] It should be noted that due to the influence of factors such as light, the grayscale values ​​of the knee joint in each area of ​​the knee joint image are different. It is difficult to segment the complete knee joint area using the same threshold. Therefore, the knee joint grayscale image needs to be processed in blocks. However, due to the small number of pixels in each block, the segmentation threshold of each block obtained using statistical features is not accurate enough.

[0051] It should be further explained that, due to the larger grayscale value of knee joint pixels, knee joint pixels require a larger segmentation threshold than non-knee joint pixels. Therefore, when different threshold segmentation is used, the knee joint pixels have a stronger ability to exist in the binary image under different thresholds; therefore, the retention rate of each pixel is obtained based on this feature.

[0052] Specifically, the knee joint grayscale image is evenly divided into image blocks, this embodiment uses 40000 is taken as an example for description, and other embodiments may take other values, which are not limited in this embodiment.

[0053] Furthermore, for any image block, the Otsu threshold algorithm is used to obtain the baseline segmentation threshold of the image block. , set the variance adjustment variable , Gaussian variance , let k take all integers between [0,20] to obtain multiple Gaussian variances, b represents the preset adjustment step size, and each adjusted segmentation threshold is obtained based on each Gaussian variance: ,in Indicates adjusting the segmentation threshold. Represents an exponential function with a natural constant as the base. The grayscale values ​​of pixels in the image block whose grayscale values ​​are greater than the adjusted segmentation threshold are set to 1, and the grayscale values ​​of pixels whose grayscale values ​​are less than or equal to the adjusted segmentation threshold are set to 0, thereby obtaining a binary image block of the image block under each adjusted segmentation threshold.

[0054] It should be noted that the variance of each Gaussian function is taken as each Gaussian variance, and the mean of the Gaussian function is taken as the benchmark segmentation threshold to obtain multiple Gaussian functions with different variances. The calculation formula for the adjusted segmentation threshold under each Gaussian variance is the Gaussian function value when the Gaussian function of each variance takes 0.

[0055] Similarly, each image block is segmented using each adjusted segmentation threshold to obtain a binary image block of each image block under each adjusted segmentation threshold.

[0056] The adjusted segmentation thresholds are arranged in ascending order according to the size of the Gaussian variance to obtain an adjusted segmentation threshold sequence.

[0057] Furthermore, for any image block, any adjusted segmentation threshold is recorded as the first target segmentation threshold, the pixel at any position is recorded as the target pixel, the grayscale value at the target pixel in the binary image block of the first target segmentation threshold is recorded as the first grayscale value of the target pixel, and the grayscale value at the target pixel in the binary image block of the previous adjusted segmentation threshold is recorded as the second grayscale value of the target pixel. When the first grayscale value is equal to 1 and the second grayscale value is equal to 0, the Gaussian variance corresponding to the first target segmentation threshold is used as the retention rate of the target pixel to obtain the retention rate of each pixel in the image block.

[0058] Similarly, the retention rate of each pixel in each image block is obtained.

[0059] So far, the retention rate of each pixel is obtained through the above method.

[0060] The retention degree acquisition module of each image block obtains the retention degree of each image block according to the retention rate of each pixel.

[0061] It should be noted that when the value of the adjusted segmentation threshold changes in the same way, the grayscale value changes of the corresponding binary image blocks are different. For example, the difference between any two adjacent adjusted segmentation thresholds is the same, but the grayscale changes of the binary image blocks corresponding to each two adjacent segmentation thresholds are different. Sometimes the change in the adjusted segmentation threshold will cause a large change in the grayscale value in the binary image block, and sometimes the change in the adjusted segmentation threshold will cause a small change in the grayscale value in the binary image block. Therefore, the change weight of each adjusted segmentation threshold is different. Therefore, the retention degree of each image block is obtained by combining the change weight and the retention rate of each pixel.

[0062] Specifically, for any image block, any adjusted segmentation threshold is recorded as the second target segmentation threshold, the number of pixels with a grayscale value of 1 in the binary image block under the second target segmentation threshold is recorded as the first number of the second target segmentation threshold, the number of pixels with a grayscale value of 1 in the binary image block of the previous adjusted segmentation threshold is recorded as the second number of the second target segmentation threshold, and the difference between the first number and the second number is recorded as the change difference of the second target segmentation threshold; the change difference of each adjusted segmentation threshold is obtained, the Gaussian variance with the same retention rate as each pixel is recorded as the benchmark Gaussian variance, the adjusted segmentation threshold corresponding to the benchmark Gaussian variance is used as the benchmark adjusted segmentation threshold of each pixel, and the change difference of the benchmark adjusted segmentation threshold of each pixel is used as the change weight of each pixel. The larger the value, the more pixels have grayscale value changes caused by the change of the adjusted segmentation threshold, and thus the change of the adjusted segmentation threshold has a greater impact.

[0063] Similarly, the change weight of each pixel in each image block is obtained.

[0064] Furthermore, the degree of retention of each image block is obtained according to the retention rate and change weight of each pixel in each image block:

[0065]

[0066] in, It represents the retention rate of the ith pixel in the zth image block. The larger the value, the larger the adjustment segmentation threshold that causes the gray value change at the pixel in the binary image block. Therefore, the gray value at the pixel in the image block is larger, and the possibility that the pixel is a knee joint pixel is greater. It represents the change weight of the ith pixel in the zth image block. The larger the value, the greater the influence of the adjustment segmentation threshold on the change of the gray value of the pixel in the binary image block. n represents the number of pixels in each image block. Indicates the degree of retention of the zth image block. The larger the value, the greater the possibility that the image block is the knee joint area.

[0067] So far, the preservation degree of each image block is obtained through the above process.

[0068] The evaluation value acquisition module for each adjusted segmentation threshold obtains the evaluation value of each adjusted segmentation threshold according to the retention degree of each image block and the grayscale value of the binary image block under each adjusted segmentation threshold.

[0069] It should be noted that, due to the dense distribution of knee joint pixels, when the retention degree of an image block is large, the grayscale values ​​of most pixels in the image block are large, and thus the image block is more likely to be the knee joint area. At the same time, for each binary image block corresponding to the adjusted segmentation threshold, when the size of the sum of the grayscale values ​​of all pixels in the binary image block is more consistent with the retention degree of the image block, the adjusted segmentation threshold is more accurate, and when the size of the sum of the grayscale values ​​of all pixels in the binary image block is more inconsistent with the retention degree of the image block, the adjusted segmentation threshold is less accurate, and thus the evaluation value of each adjusted segmentation threshold is obtained based on this.

[0070] Specifically, for any image block, the gray value mean of all pixels in the binary image block under each adjusted segmentation threshold is recorded as the gray value mean of the image block under each adjusted segmentation threshold, and the gray value mean of the image block under each adjusted segmentation threshold is normalized using the maximum and minimum normalization method to obtain the normalized gray value mean of the image block under each adjusted segmentation threshold; the retention degree of the image block is normalized using the maximum and minimum normalization method to obtain the normalized retention degree of the image block. For the sake of ease of description, the normalized gray value mean of the image block under each adjusted segmentation threshold is still recorded as the gray value mean of the image block under each adjusted segmentation threshold, and the normalized retention degree of the image block is still recorded as the retention degree of the image block. The evaluation value of each adjusted segmentation threshold is obtained according to the retention degree of each image block and the binary image under each adjusted segmentation threshold of each image block:

[0071]

[0072] in, Indicates the degree of preservation of the zth image block. The larger the value, the greater the possibility that the zth image block is the knee joint area. It represents the mean gray value of the zth image block under each adjusted segmentation threshold. The larger the value is, the more likely it is that the image block is the knee joint area under the adjusted segmentation threshold. exp() represents an exponential function with a natural constant as the base. It reflects the consistency between the retention degree of the z-th image block and the mean gray value under each adjusted segmentation threshold. The larger the value, the better the segmentation effect of the adjusted segmentation threshold. represents the number of image blocks, and P represents the evaluation value of each adjusted segmentation threshold.

[0073] So far, the evaluation value of each adjusted segmentation threshold is obtained through the above process.

[0074] The image segmentation module takes the adjusted segmentation threshold with the largest evaluation value as the optimal segmentation threshold, and uses the optimal segmentation threshold to segment the knee joint grayscale image to obtain the knee joint area.

[0075] Specifically, the adjusted segmentation threshold with the largest evaluation value is obtained as the optimal segmentation threshold, the grayscale value of pixels in the knee joint grayscale image that are greater than the optimal segmentation threshold is set to 1, and the grayscale value of pixels in the knee joint grayscale image that are less than or equal to the optimal segmentation threshold is set to 0 to obtain the optimal binary image, where pixels with grayscale values ​​equal to 1 in the optimal binary image are knee joint pixels.

[0076] The connected domain formed by knee joint pixels is regarded as the knee joint area.

[0077] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. The intelligent segmentation system of knee joint images based on local threshold is characterized by: The system includes the following modules: A knee joint grayscale image acquisition module, used for acquiring a knee joint grayscale image; A retention rate acquisition module for each pixel is used to process the knee joint grayscale image into blocks to obtain multiple image blocks, set multiple adjustment segmentation thresholds for each image block, use each adjustment segmentation threshold to perform segmentation processing on each image block to obtain a binary image block of each image block under each adjustment segmentation threshold, and obtain the retention rate of each pixel according to the change of the grayscale value of each pixel in two binary image blocks under two adjacent adjustment segmentation thresholds; A retention degree acquisition module for each image block, used to obtain the change weight of each pixel according to the number of gray value change pixels in two adjacent binary image blocks under the adjusted segmentation threshold, and obtain the retention degree of each image block according to the change weight of each pixel and the retention rate of each pixel; An evaluation value acquisition module for each adjusted segmentation threshold, used to obtain an evaluation value for each adjusted segmentation threshold according to the degree of retention of each image block and the consistency of the mean grayscale value in the binary image block of each image block under each adjusted segmentation threshold; An image segmentation module is used to obtain an optimal segmentation threshold according to the evaluation value of each adjusted segmentation threshold, and to segment the knee joint grayscale image according to the optimal segmentation threshold to obtain a knee joint area; The specific method of setting multiple adjustment segmentation thresholds for each image block includes: For any image block, the Otsu threshold algorithm is used to obtain the baseline segmentation threshold of the image block. , set the variance adjustment variable , Gaussian variance , let k take all integers between [0,20] to obtain multiple Gaussian variances, b represents the preset adjustment step size, and each adjusted segmentation threshold is obtained based on each Gaussian variance: ,in Indicates adjusting the segmentation threshold. represents an exponential function with a natural constant as base; Wherein, the retention rate of each pixel is obtained according to the change of the gray value of each pixel in two binary image blocks under two adjacent adjusted segmentation thresholds, and the specific method includes: arranging the adjusted segmentation thresholds in ascending order according to the size of the Gaussian variance to obtain an adjusted segmentation threshold sequence; for any image block, recording any adjusted segmentation threshold as the first target segmentation threshold, recording a pixel at any position as a target pixel, recording the gray value at the target pixel in the binary image block of the first target segmentation threshold as the first gray value of the target pixel, recording the gray value at the target pixel in the binary image block of the previous adjusted segmentation threshold as the second gray value of the target pixel, and when the first gray value is equal to 1 and the second gray value is equal to 0, taking the Gaussian variance corresponding to the first target segmentation threshold as the retention rate of the target pixel, and obtaining the retention rate of each pixel in the image block; Among them, the change weight of each pixel is obtained according to the number of pixels with changed grayscale values ​​in two binary image blocks under two adjacent adjusted segmentation thresholds, and the specific method includes: for any image block, any adjusted segmentation threshold is recorded as the second target segmentation threshold, the number of pixels with a grayscale value of 1 in the binary image block under the second target segmentation threshold is recorded as the first number of the second target segmentation threshold, the number of pixels with a grayscale value of 1 in the binary image block of the previous adjusted segmentation threshold is recorded as the second number of the second target segmentation threshold, and the difference between the first number and the second number is recorded as the change difference of the second target segmentation threshold; the change difference of each adjusted segmentation threshold is obtained, the Gaussian variance with the same retention rate as each pixel is recorded as the base Gaussian variance, the adjusted segmentation threshold corresponding to the base Gaussian variance is used as the base adjusted segmentation threshold of each pixel, and the change difference of the base adjusted segmentation threshold of each pixel is used as the change weight of each pixel; the change weight of each pixel in each image block is obtained.

2. According to claim 1, the local threshold-based intelligent segmentation system for knee joint images is characterized in that: The grayscale image of the knee joint is processed into blocks to obtain a plurality of image blocks, and the specific method includes: The knee joint grayscale image is evenly divided into multiple image blocks.

3. The local threshold-based intelligent segmentation system for knee joint images according to claim 1 is characterized in that: The specific method of obtaining the retention degree of each image block according to the change weight of each pixel and the retention rate of each pixel includes: in, represents the retention rate of the i-th pixel in the z-th image block; represents the change weight of the i-th pixel in the z-th image block, n represents the number of pixels in each image block, Indicates the degree of preservation of the z-th image block.

4. The local threshold-based intelligent knee joint image segmentation system according to claim 1, characterized in that: The specific method of obtaining the evaluation value of each adjusted segmentation threshold according to the degree of retention of each image block and the consistency of the mean grayscale value in the binary image block of each image block under each adjusted segmentation threshold is as follows: The mean gray value of all pixels in the binary image block under each adjusted segmentation threshold is recorded as the mean gray value of the image block under each adjusted segmentation threshold; The calculation method for obtaining the evaluation value of each adjusted segmentation threshold according to the mean gray value of each image block at each adjusted segmentation threshold and the retention degree of each image block is: in, Indicates the degree of preservation of the z-th image block, represents the mean gray value of the z-th image block under each adjusted segmentation threshold, represents the number of image blocks, P represents the evaluation value of each adjusted segmentation threshold, and exp() represents an exponential function with a natural constant as the base.

5. The local threshold-based intelligent knee joint image segmentation system according to claim 1, characterized in that: The specific method of obtaining the optimal segmentation threshold according to the evaluation value of each adjusted segmentation threshold is as follows: The adjusted segmentation threshold with the largest evaluation value is taken as the optimal segmentation threshold.

6. The local threshold-based intelligent segmentation system for knee joint images according to claim 1, characterized in that: The specific method of performing segmentation processing on each image block using each adjusted segmentation threshold to obtain a binary image block of each image block under each adjusted segmentation threshold is as follows: The grayscale values ​​of pixels in each image block whose grayscale values ​​are greater than each adjusted segmentation threshold are set to 1, and the grayscale values ​​of pixels whose grayscale values ​​are less than or equal to each adjusted segmentation threshold are set to 0, thereby obtaining a binary image block of each image block under each adjusted segmentation threshold.

7. The local threshold-based intelligent segmentation system for knee joint images according to claim 1, characterized in that: The method of segmenting the knee joint grayscale image according to the optimal segmentation threshold to obtain the knee joint area includes: The grayscale values ​​of pixels in the knee joint grayscale image that are greater than the optimal segmentation threshold are set to 1, and the grayscale values ​​of pixels in the knee joint grayscale image that are less than or equal to the optimal segmentation threshold are set to 0 to obtain an optimal binary image, wherein the pixels in the optimal binary image whose grayscale values ​​are equal to 1 are knee joint pixels; The connected domain formed by knee joint pixels is regarded as the knee joint area.

Images