Knee joint MR image enhancement method and system

Abstract

The invention relates to the technical field of image data processing, in particular to a knee joint MR image enhancement method and system. The method comprises the following steps: according to the gradient direction similarity of pixel points in a knee joint MR image between different scales and the difference between the amplitude value of the pixel points and the maximum amplitude value in the knee joint MR image, determining a noise map and carrying out separation to obtain a target knee joint MR image; according to the ratio of the area of a super-pixel block in the target knee joint MR image to the area of the minimum enclosing rectangle of the super-pixel block, the difference between the gray average value and the maximum gray value of the super-pixel block, and the gray average value difference between the super-pixel block and peripheral pixel points, a cartilage collagen fiber region is obtained; in the Laplace operator, the cartilage collagen fiber area and other areas in the target knee joint MR image are sharpened and enhanced by using different coefficients, the enhanced knee joint MR image is obtained, and the accuracy of knee joint MR image enhancement is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method and system for enhancing MR images of a knee joint. Background Art

[0002] Knee osteoarthritis is a degenerative disease. Before using platelet-rich plasma to treat knee osteoarthritis, it is necessary to use knee magnetic resonance imaging (MR) to examine the knee cartilage. MR can clearly display the morphology, structure and pathological conditions of soft tissue, providing efficient and objective diagnostic support for clinical practice.

[0003] During knee MR imaging, radiofrequency pulses excite hydrogen nuclei within the human body and receive the radiofrequency signals emitted by these nuclei, generating a two-dimensional intensity matrix. The value of each pixel in the knee MR image represents the signal strength at that location. Higher signal strengths correspond to brighter pixels, while lower signal strengths correspond to darker pixels. To reduce the blurring of MR image details caused by the complex structure of the knee joint, knee MR images are sharpened and enhanced using the Laplacian operator. This enhances image contrast and clarity by highlighting high-frequency details (such as edges and textures).

[0004] However, the Laplace operator is a global edge enhancement algorithm that cannot accurately distinguish between useful high-frequency areas that need to be identified (such as the boundary between cartilage and bone) and useless low-frequency areas that do not need to be enhanced (such as randomly distributed colored noise points generated by random thermal noise of the device and the microscopic arrangement of cartilage collagen fibers). As a result, useless low-frequency areas are also enhanced, causing misjudgment.

[0005] Therefore, how to enhance only the useful high-frequency areas on knee joint MR images when using the Laplacian operator to enhance images is a problem that needs to be solved urgently. Summary of the Invention

[0006] In order to solve the technical problem of how to enhance only the useful high-frequency region on a knee joint MR image when using a Laplacian operator to enhance an image, the present invention provides a knee joint MR image enhancement method and system.

[0007] In a first aspect, the present invention provides a method for enhancing MR images of a knee joint, which employs the following technical solutions: A method for enhancing MR images of a knee joint comprises the following steps: The method comprises the following steps: obtaining the gradient direction of each pixel in the knee joint MR image at different scales; obtaining the degree to which the pixel belongs to noise according to the similarity of the gradient direction between the pixel at different scales and the difference between the amplitude value of the pixel and the maximum amplitude value in the knee joint MR image; generating a noise map in response to the comparison result of the degree to which the pixel belongs to noise and the noise threshold, separating the noise map from the knee joint MR image, and obtaining a target knee joint MR image; dividing the target knee joint MR image into multiple super-pixel blocks, determining the area of each super-pixel block; and determining the noise level of the super-pixel block according to the difference between the area of the super-pixel block and the maximum amplitude value of the super-pixel block. The probability that the superpixel block belongs to cartilage collagen fibers is obtained by comparing the probability that the superpixel block belongs to cartilage collagen fibers with a probability threshold. The cartilage collagen fiber region is obtained based on the grayscale mean difference between the candidate cartilage collagen fiber region and its surrounding pixels. Different coefficients are used in the Laplace operator to perform sharpening enhancement on the cartilage collagen fiber region and other regions in the target knee joint MR image to obtain an enhanced knee joint MR image.

[0008] The present invention uses the Laplace operator to sharpen and enhance the knee joint MR image, which can highlight the details in the image and improve the image clarity. During the image enhancement process, the present invention takes into account that different areas in the knee joint MR image have different enhancement requirements, such as cartilage collagen fibers; therefore, the present invention can accurately calculate and identify the cartilage collagen fiber area and other areas in the image by analyzing the grayscale characteristics, brightness characteristics and local contrast characteristics of the superpixel blocks in the image, so that accurate enhancement can be performed according to the sharpening requirements of different areas, effectively improving the accuracy of knee joint MR image enhancement. Before identifying the cartilage collagen fiber area, the present invention also takes into account that the noise points in the knee joint MR image will affect the accuracy of cartilage collagen fiber area identification and may affect subsequent diagnosis; therefore, the present invention accurately removes the noise points in the knee joint MR image by analyzing the degree to which the amplitude characteristics and gradient characteristics of each pixel point in the image conform to the noise characteristics, thereby effectively improving the accuracy and efficiency of knee joint MR image enhancement.

[0009] According to a knee joint MR image enhancement method provided by the present invention, the method of obtaining the gradient direction of each pixel point in the knee joint MR image at different scales also includes: converting the knee joint MR image into a spectrum diagram through Fourier transform to obtain the amplitude value of each pixel point and the maximum amplitude value in the knee joint MR image.

[0010] According to a knee joint MR image enhancement method provided by the present invention, obtaining the gradient direction of each pixel point in the knee joint MR image at different scales includes: using different scales in the Sobel operator to obtain the gradient direction of each pixel point at different scales.

[0011] According to a knee joint MR image enhancement method provided by the present invention, the degree to which a pixel belongs to noise is obtained based on the gradient direction similarity between pixels at different scales and the difference between the amplitude value of the pixel and the maximum amplitude value in the knee joint MR image, including: recording the mean cosine similarity between the gradient direction of each pixel at one scale and the gradient direction of the next scale as the similarity index of the pixel; calculating the degree to which the i-th pixel belongs to noise : ; 、 are the amplitude value and similarity index of the i-th pixel, respectively. is the maximum amplitude value in the knee joint MR image, is an exponential function with base e.

[0012] The present invention takes into account that the noise points in the knee joint MR images appear as high-frequency signals in the frequency domain and appear as chaotic and inconsistent at different scales. Therefore, by analyzing the amplitude of each pixel point and the similarity of the gradient direction at different scales, the degree to which each pixel point belongs to noise is accurately calculated.

[0013] According to a knee joint MR image enhancement method provided by the present invention, the noise map is generated in response to the comparison result of the degree to which pixel points belong to noise and the noise threshold, including: taking pixel points in the knee joint MR image whose degree of noise belongs to the image is greater than the noise threshold as noise points; and generating a noise map based on the noise points in the knee joint MR image.

[0014] According to a knee joint MR image enhancement method provided by the present invention, the target knee joint MR image is divided into multiple superpixel blocks, and the area of each superpixel block is determined, including: determining a seed point in the target knee joint MR image; performing region growing with the seed point as the starting point, and finally dividing the target knee joint MR image into multiple superpixel blocks; and using the number of pixels in the superpixel block as the area of the superpixel block.

[0015] The present invention takes into account that the cartilage collagen fiber region has similar features, and superpixel blocks can group pixels with similar features together. Therefore, the present invention divides the target knee joint MR image into multiple superpixel blocks and analyzes the features within each superpixel block, thereby effectively reducing the data processing amount while fully identifying the cartilage collagen fibers.

[0016] According to a knee joint MR image enhancement method provided by the present invention, the probability that the superpixel block belongs to cartilage collagen fiber is obtained based on the ratio of the area of the superpixel block to the area of the minimum circumscribed rectangle of the superpixel block and the difference between the grayscale mean and the maximum grayscale value of the superpixel block, including: screening out superpixel blocks whose grayscale mean difference with the maximum grayscale mean of all superpixel blocks is not greater than a difference threshold; calculating the first The probability that a superpixel belongs to cartilage collagen fiber : ; For the The grayscale mean of superpixel blocks, is the maximum grayscale value, 、 Respectively The area of the superpixel block, the area of the minimum bounding rectangle, is an exponential function with base e.

[0017] The present invention takes into account that the external features of the cartilage collagen fiber area are close to a rectangle, and the grayscale features are overall darker than other areas. Therefore, by analyzing the grayscale features inside each superpixel block and its rectangularity, the possibility of each superpixel block being a cartilage collagen fiber is accurately evaluated, preparing for the subsequent accurate screening of cartilage collagen fibers.

[0018] According to a knee joint MR image enhancement method provided by the present invention, the cartilage collagen fiber region is obtained based on the grayscale mean difference between the candidate cartilage collagen fiber region and its surrounding pixel points, including: recording the normalized value of the grayscale mean difference between the candidate cartilage collagen fiber region and its surrounding pixel points as the grayscale contrast of the candidate cartilage collagen fiber region; if the grayscale contrast of any candidate cartilage collagen fiber region is not less than 0, then sorting the grayscale contrasts of the candidate cartilage collagen fiber regions with grayscale contrast not less than 0, and selecting the candidate cartilage collagen fiber region with grayscale contrast greater than or equal to the third quartile as the cartilage collagen fiber region.

[0019] The present invention takes into account that the cartilage collagen fiber region also has an important feature: compared with the surrounding area, it has a stronger grayscale contrast and is relatively brighter than the surrounding cartilage matrix. Therefore, the present invention accurately obtains the grayscale contrast of each candidate cartilage collagen fiber region by analyzing the grayscale difference between the candidate cartilage collagen fiber region and its peripheral pixel points, so that the cartilage collagen fiber region in the target knee joint MR image can be accurately obtained based on this feature.

[0020] According to a knee joint MR image enhancement method provided by the present invention, different coefficients are used in the Laplace operator to perform sharpening enhancement on the cartilage collagen fiber area and other areas in the target knee joint MR image, including: using the cartilage collagen fiber area as a low-sharpening mask, recording the area after removing the low-sharpening mask in the knee joint MR image edge map as a high-sharpening mask, and using the remaining area except the low-sharpening mask and the high-sharpening mask as a medium-sharpening mask; using a low sharpening coefficient, a medium sharpening coefficient and a high sharpening coefficient in the Laplace operator respectively to perform sharpening enhancement on the low-sharpening mask, the medium sharpening coefficient and the high-sharpening mask; wherein the low sharpening coefficient is smaller than the medium sharpening coefficient, and the medium sharpening coefficient is smaller than the high sharpening coefficient.

[0021] In a second aspect, the present invention provides a knee joint MR image enhancement system, which adopts the following technical solution: A knee joint MR image enhancement system includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned knee joint MR image enhancement method is implemented.

[0022] By adopting the above technical solution, the above-mentioned knee joint MR image enhancement method is generated into a computer program and stored in a memory to be loaded and executed by a processor, thereby making a terminal device based on the memory and the processor for easy use.

[0023] The present invention has the following technical effects: Based on the above technical solution, the present invention provides a knee joint MR image enhancement method and system, which can sharpen and enhance the knee joint MR image through the Laplace operator, so as to highlight the details in the image and improve the image clarity. During the image enhancement process, the present invention can accurately calculate and identify the cartilage collagen fiber area and other areas in the image by analyzing the grayscale characteristics, brightness characteristics and local contrast characteristics of the super-pixel blocks in the image, so as to accurately enhance the sharpening requirements of different areas, effectively improving the accuracy of knee joint MR image enhancement. Before identifying the cartilage collagen fiber area, the present invention also accurately removes the noise in the knee joint MR image by analyzing the degree to which the amplitude characteristics and gradient characteristics of each pixel point in the image conform to the noise characteristics, thereby effectively improving the accuracy and efficiency of knee joint MR image enhancement. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic diagram of the flow of a knee joint MR image enhancement method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0026] In order to enhance only the useful high-frequency areas on the knee joint MR image when using the Laplacian operator to enhance the image, an embodiment of the present invention discloses a knee joint MR image enhancement method. The method analyzes the possibility that each superpixel block in the knee joint MR image belongs to the cartilage collagen fiber area, obtains the cartilage collagen fiber area and other areas in the knee joint MR image, so that the corresponding sharpening coefficient can be accurately set based on this, thereby effectively improving the accuracy of knee joint MR image enhancement.

[0027] For details, please see Figure 1 As shown, Figure 1 This is a flow chart of a method for enhancing MR images of a knee joint provided by an embodiment of the present invention. The method specifically includes the following steps: S1: Obtain the grayscale value, amplitude value and gradient direction of each pixel in the knee joint MR image at different scales.

[0028] It should be noted that when acquiring knee MR images, random thermal noise inside the acquisition device will cause noise interference to the knee MR images. The noise in the knee MR images is not only a useless low-frequency area that does not need to be enhanced, but also affects the subsequent diagnostic results.

[0029] Based on this, in an embodiment of the present invention, before identifying the cartilage collagen fiber region in the knee joint MR image, the knee joint MR image may be denoised first, while retaining the image details and clear edges during the denoising process.

[0030] It can be understood that noise appears as a high-frequency signal in the frequency domain, and the gradient direction of the edge of the knee joint in the knee MR image is usually consistent at different scales, while noise is random and independent, and its gradient direction will appear chaotic at different scales.

[0031] Therefore, the embodiment of the present invention can accurately calculate the possibility that each pixel is noise by combining the amplitude value of each pixel to identify the high-frequency noise component in the image and the gradient direction consistency of the pixel at different scales.

[0032] For example, in an embodiment of the present invention, obtaining the grayscale value and amplitude value of each pixel point in the knee joint MR image includes: converting the knee joint MR image into a spectrum diagram through Fourier transform to obtain the amplitude value of each pixel point and the maximum amplitude value in the knee joint MR image.

[0033] Specifically, the MRI3d-Dess cartilage scanning technology can be used to collect MR images of the knee joint and grayscale processing can be performed to obtain the knee joint MR image. The grayscale value range of the pixel points in the knee joint MR image is [0, 255]; the parameters during the scanning process can be set to: field of view 160mm, layer thickness 3mm, repetition time 3900ms, echo time 43ms, matrix 256×320; the scanning parameters can be set according to actual needs, and the embodiment of the present invention does not impose too many restrictions here.

[0034] For example, when converting the knee joint MR image into a spectrum diagram through Fourier transform to obtain the amplitude value of each pixel point and the maximum amplitude value in the knee joint MR image, the spectrum diagram can be constructed with frequency as the horizontal axis and the amplitude value corresponding to each frequency as the vertical axis to obtain the amplitude value of each pixel point; the maximum amplitude value in the spectrum diagram is obtained as the maximum amplitude value in the knee joint MR image.

[0035] The step of converting the knee joint MR image into a spectrum diagram through Fourier transform can be obtained through existing technology and will not be described in detail in the embodiment of the present invention.

[0036] For example, in an embodiment of the present invention, obtaining the gradient direction of each pixel point in a knee joint MR image at different scales includes: using different scales in a Sobel operator to obtain the gradient direction of each pixel point at different scales.

[0037] Among them, in the Sobel operator, the scale is determined by the size of the Gaussian kernel. The size corresponding to each scale can be obtained through the formula of the scale sequence, which usually increases in the form of a high-order power of 2.

[0038] In this embodiment of the present invention, the number of scales can be set to 3, with the initial value of the scale being 0, resulting in the corresponding sizes of 1, 2, and 4, respectively. This number can be set based on actual needs. The formula for the scale sequence and the specific steps for obtaining the gradient direction of a pixel at each scale can be obtained from existing technologies and are not detailed in this embodiment of the present invention.

[0039] After obtaining the amplitude value and gradient direction of each pixel point based on the above steps, the possibility of each pixel point being noise data can be accurately obtained by combining the frequency characteristics and gradient characteristics, thereby accurately identifying the noise points in the knee joint MR image, that is, performing the following steps.

[0040] S2: The degree to which a pixel belongs to noise is obtained based on the similarity of the gradient direction of the pixel at different scales and the difference between the amplitude value of the pixel and the maximum amplitude value in the knee joint MR image.

[0041] It's important to note that noise typically manifests as random, irregular fluctuations in images. Therefore, in knee MR images, the maximum amplitude values often represent extreme noise rather than a true representation of tissue signal. The smaller the difference between each pixel and the maximum amplitude value in a knee MR image, the more chaotic the gradient directions of the pixel at different scales, and the lower the similarity, the more likely that the pixel is noise.

[0042] Based on this, the embodiment of the present invention can accurately obtain the degree to which a pixel point belongs to noise based on the gradient direction similarity between pixels at different scales and the difference between the amplitude value of the pixel point and the maximum amplitude value in the knee joint MR image.

[0043] For example, in an embodiment of the present invention, the degree to which a pixel point belongs to noise is obtained based on the similarity of the gradient directions of the pixel points between different scales, and the difference between the amplitude value of the pixel point and the maximum amplitude value in the knee joint MR image. This also includes: recording the average cosine similarity of each pixel point between the gradient direction of one scale and the gradient direction of the next scale as the similarity index of the pixel point.

[0044] Among them, the similarity index is used to characterize the gradient direction consistency of pixels at different scales.

[0045] An example is given to illustrate how to obtain the similarity index of a pixel: the number of scales is 3, numbered 0, 1, and 2 respectively. The cosine similarities between the gradient directions of scale 0 and scale 1, scale 1 and scale 2, and scale 2 and scale 3 are calculated respectively; the average of the three groups of cosine similarities is used as the similarity index of the pixel.

[0046] For example, to calculate the degree to which a pixel point belongs to noise, please refer to the following relationship: ; is the degree to which the i-th pixel belongs to noise, is the amplitude value of the i-th pixel, is the similarity index of the i-th pixel, is the maximum amplitude value in the knee joint MR image, It is an exponential function with base e, where e is a natural constant.

[0047] In the above formula, is the normalized representation of the frequency domain feature of the i-th pixel. If the difference between the amplitude value of the i-th pixel and the maximum amplitude value is The smaller it is, the greater the possibility that the i-th pixel belongs to noise in the frequency domain.

[0048] Indicates the consistency of the gradient direction of the i-th pixel at different scales. The smaller the value, the greater the degree of confusion of the gradient direction of the i-th pixel at different scales in the spatial domain, the smaller the similarity, and the greater the possibility of being noise.

[0049] In summary, the smaller the similarity of the gradient direction of the pixel point between different scales and the smaller the difference between the amplitude value of the pixel point and the maximum amplitude value in the knee joint MR image, the greater the possibility that the pixel point is noise.

[0050] Thus, by combining frequency and gradient features, the present invention can more comprehensively identify different types of noise, including high-frequency noise and random noise caused by the device, thereby avoiding misclassifying normal texture details as noise and preserving useful information in the image.

[0051] After obtaining the degree to which each pixel point belongs to noise based on the above steps, the noise points in the knee joint MR image can be accurately determined according to the degree to which the pixel point belongs to noise.

[0052] S3: In response to the comparison result of the degree to which the pixel point belongs to noise and the noise threshold, a noise map is generated, the noise map is separated from the knee joint MR image, and a target knee joint MR image is obtained; the target knee joint MR image is divided into multiple superpixel blocks, and the area of each superpixel block is determined.

[0053] The noise threshold may be set to 0.6; the noise threshold may be set specifically according to actual needs.

[0054] For example, in an embodiment of the present invention, a noise map is generated in response to a comparison result of the degree to which a pixel point belongs to noise and a noise threshold, including: treating pixel points in a knee joint MR image whose degree to which a pixel point belongs to noise is greater than the noise threshold as noise points; and generating a noise map based on the noise points in the knee joint MR image.

[0055] It can be understood that if the degree to which a pixel point in the knee joint MR image belongs to noise is not greater than the noise threshold, it means that such pixel points do not conform to the frequency domain characteristics and gradient characteristics of noise and do not need to be removed. The value of each pixel point in the noise map is the degree to which each pixel point belongs to noise.

[0056] After separating the noise map from the knee joint MR image based on the above steps, a target knee joint MR image that is not interfered by noise can be obtained. In the target knee joint MR image, the cartilage collagen fiber area has specific morphological and textural characteristics. The superpixel block can group pixels with similar characteristics together to form a more meaningful regional unit. When the cartilage collagen fiber area is subsequently analyzed, the superpixel block can accurately capture the boundaries and internal features of these areas, avoiding the information fragmentation problem that may occur when analyzing based on a single pixel point, and helping to more completely identify the structure of the cartilage collagen fibers. Therefore, the embodiment of the present invention can divide the target knee joint MR image into multiple superpixel blocks, analyze each superpixel block, and accurately screen out superpixel blocks that meet the characteristics of the cartilage collagen fibers.

[0057] For example, in an embodiment of the present invention, the target knee joint MR image is divided into multiple superpixel blocks, and the area of each superpixel block is determined, including: determining a seed point in the target knee joint MR image; performing region growing with the seed point as the starting point, and finally dividing the target knee joint MR image into multiple superpixel blocks; and taking the number of pixels in the superpixel block as the area of the superpixel block.

[0058] The size of the seed point may be set to 3×3. The size of the seed point may be set according to actual needs, and the embodiment of the present invention does not impose any additional restrictions thereon.

[0059] Specifically, a 3×3 pixel block is selected as a seed point on the target knee joint MR image every 4 pixels. For each 3×3 seed point, its grayscale variance is calculated; the seed point with a grayscale value variance greater than a preset threshold is selected as the starting point for region growing, and finally multiple superpixel blocks are obtained.

[0060] The preset threshold value can be set according to actual needs.

[0061] After the target knee joint MR image is divided into multiple superpixel blocks based on the above steps, each superpixel block can be accurately analyzed to accurately obtain the probability that it belongs to cartilage collagen fibers, that is, perform the following steps.

[0062] S4: According to the ratio of the area of the super pixel block to the area of the minimum circumscribed rectangle of the super pixel block and the difference between the grayscale mean value and the maximum grayscale value of the super pixel block, the probability that the super pixel block belongs to the cartilage collagen fiber is obtained.

[0063] It should be noted that cartilage collagen fibers are linear or ribbon-like structures, which appear as elongated, continuous lines in MR images, approximately rectangular. Furthermore, there is typically high grayscale contrast between collagen fibers and the surrounding cartilage matrix, with a clear boundary. Therefore, collagen fibers appear as relatively bright areas in the image, while the surrounding cartilage matrix is darker. The grayscale value has a distinct transition at the boundary, which appears as a distinct line edge in MR images.

[0064] Based on this, the embodiment of the present invention can accurately obtain the probability that the superpixel block belongs to cartilage collagen fibers based on the ratio of the area of the superpixel block to the area of the minimum circumscribed rectangle of the superpixel block, and the difference between the grayscale mean and the maximum grayscale value of the superpixel block.

[0065] It should be further explained that the superpixel blocks obtained based on the above steps include both cartilage collagen fiber regions and residual regions with higher brightness. The residual regions with higher brightness are clearly non-cartilage collagen fiber regions. Furthermore, the superpixel blocks obtained based on region growing are irregular in size and have low rectangularity, which clearly does not conform to the brightness and rectangular characteristics of the cartilage collagen fiber regions. Therefore, before calculating the probability that each superpixel block belongs to cartilage collagen fibers based on the following steps, non-cartilage collagen fiber regions that clearly do not conform to the cartilage collagen fiber characteristics can be screened out, thereby effectively improving data processing efficiency and reducing data processing volume.

[0066] For example, in an embodiment of the present invention, the probability that the superpixel block belongs to cartilage collagen fibers is obtained based on the ratio of the area of the superpixel block to the area of the minimum circumscribed rectangle of the superpixel block and the difference between the grayscale mean and the maximum grayscale value of the superpixel block. This also includes: screening out superpixel blocks whose difference between the grayscale mean in the superpixel block and the maximum grayscale mean of all superpixel blocks is not greater than a difference threshold.

[0067] The difference threshold may be set to 0.5. The difference threshold may be set according to actual needs, and the embodiment of the present invention does not impose any additional restrictions thereon.

[0068] Specifically, when obtaining the difference between the grayscale mean in the superpixel block and the maximum grayscale mean of all superpixel blocks, the absolute value of the difference between the grayscale mean in the superpixel block and the maximum grayscale mean of all superpixel blocks can be normalized to obtain the difference between the grayscale mean in the superpixel block and the maximum grayscale mean of all superpixel blocks. After screening out non-cartilage collagen fiber superpixel blocks that obviously do not meet the characteristics of cartilage collagen fibers based on the difference, continue to perform the following steps to calculate the probability of each superpixel block belonging to cartilage collagen fibers.

[0069] For example, in an embodiment of the present invention, the probability that a superpixel block belongs to a cartilage collagen fiber is calculated, and the specific relationship can be referred to as follows: ; For the The probability that a superpixel block belongs to cartilage collagen fiber, For the The grayscale mean of superpixel blocks, is the maximum grayscale value, For the The area of a superpixel block, For the The minimum bounding rectangle area of a superpixel block, is an exponential function with base e.

[0070] In the above formula, Indicates the The rectangularity of the superpixel block, the closer the value is to 1, the The closer the minimum circumscribed rectangle of a superpixel block is to a rectangle, the more it conforms to the shape characteristics of cartilage collagen fibers.

[0071] Indicates the The normalized close value between the grayscale mean and the maximum grayscale value of the superpixel block. The grayscale of the cartilage collagen fiber belongs to the area with lower overall grayscale value in the image. Therefore, the smaller the value, the closer the superpixel block is to the grayscale mean and the maximum grayscale value, the closer the superpixel block is to the grayscale mean. The closer the grayscale mean of each superpixel is to the maximum grayscale value, the less it conforms to the brightness characteristics of cartilage collagen fibers. Close to 1, it does not conform to the characteristics of cartilage collagen fibers, so the corresponding The probability that the superpixel block belongs to cartilage collagen fiber is smaller. On the contrary, the larger the value, the The smaller the grayscale mean value of a superpixel block is than the maximum grayscale value, the more it conforms to the brightness characteristics of cartilage collagen fibers, and the greater the probability that it belongs to cartilage collagen fibers.

[0072] Based on the above steps, the probability that each superpixel block belongs to cartilage collagen fibers can be obtained. According to the probability that the superpixel block belongs to cartilage collagen fibers, the superpixel blocks belonging to cartilage collagen fibers in the superpixel block can be accurately obtained.

[0073] S5: In response to the comparison result of the probability that the superpixel block belongs to the cartilage collagen fiber and the probability threshold, a candidate cartilage collagen fiber region is obtained, the peripheral pixel points of the candidate cartilage collagen fiber region are determined, and the cartilage collagen fiber region is obtained according to the grayscale mean difference between the candidate cartilage collagen fiber region and its peripheral pixel points.

[0074] For example, when setting the probability threshold, the probabilities of all superpixel blocks belonging to cartilage collagen fibers can be sorted, and the third quartile of the sorting can be used as the probability threshold.

[0075] The third quartile of the ranking may be obtained by linear interpolation. The specific steps may be obtained by existing technologies and will not be described in detail in the embodiment of the present invention.

[0076] It is understandable that, because cartilage collagen fibers have a characteristic of high local contrast, if a superpixel block corresponds to a complete cartilage collagen fiber region, the brightness within the superpixel block is higher than the brightness of its surrounding pixels. Therefore, if the brightness within the superpixel block is not less than the brightness of its surrounding pixels, the characteristics of the candidate cartilage collagen fiber region meet the local characteristics of cartilage collagen fibers; otherwise, the characteristics of the candidate cartilage collagen fiber region do not meet the local characteristics of cartilage collagen fibers.

[0077] Based on this, the real cartilage collagen fiber region can be obtained from all candidate cartilage collagen fiber regions.

[0078] For example, in an embodiment of the present invention, a cartilage collagen fiber region is obtained based on the grayscale mean difference between the candidate cartilage collagen fiber region and its surrounding pixel points, including: recording the normalized value of the grayscale mean difference between the candidate cartilage collagen fiber region and its surrounding pixel points as the grayscale contrast of the candidate cartilage collagen fiber region; if the grayscale contrast of any candidate cartilage collagen fiber region is not less than 0, then sorting the grayscale contrasts of the candidate cartilage collagen fiber regions with grayscale contrast not less than 0, and selecting the candidate cartilage collagen fiber region with grayscale contrast greater than or equal to the third quartile as the cartilage collagen fiber region.

[0079] The peripheral pixel points of the candidate cartilage collagen fiber region are a circle of pixel points adjacent to the outer circle of the candidate cartilage collagen fiber region.

[0080] It can be understood that if the grayscale contrast of any candidate cartilage collagen fiber region is not less than 0, it means that the grayscale mean of the candidate cartilage collagen fiber region is not less than the grayscale mean between its peripheral pixels, and the characteristics of the candidate cartilage collagen fiber region meet the local characteristics of cartilage collagen fibers. If the grayscale contrast of any candidate cartilage collagen fiber region is not less than 0, it means that at least one candidate cartilage collagen fiber region meets the local characteristics of cartilage collagen fibers. At this time, the grayscale contrasts of all candidate cartilage collagen fiber regions with grayscale contrasts not less than 0 can be sorted, and the candidate cartilage collagen fiber regions can be screened. On the contrary, if the grayscale contrast of the candidate cartilage collagen fiber region is less than 0, it means that it does not meet the local characteristics of cartilage collagen fibers.

[0081] In this way, the embodiment of the present invention reflects the distribution of grayscale contrast data of all candidate regions through the third quartile, so that the candidate cartilage collagen fiber regions with larger grayscale contrast can be accurately screened out from all candidate cartilage collagen fiber regions with grayscale contrast not less than 0 as cartilage collagen fiber regions, reducing the influence of other factors on the recognition results.

[0082] After obtaining the cartilage collagen fiber region in the target knee joint MR image based on the above steps, different degrees of sharpening enhancement can be performed on different regions, that is, the following steps are performed.

[0083] S6: Using different coefficients in the Laplace operator, the cartilage collagen fiber region and other regions in the target knee joint MR image are sharpened and enhanced to obtain an enhanced knee joint MR image.

[0084] It should be noted that in the target knee joint MR image, the cartilage collagen fiber area is a low-intensity enhanced low-sharpening area to avoid affecting the subsequent diagnosis; the edge in the target knee joint MR image is the real edge of the anatomical structure, which is a high-sharpening area that needs to be highly enhanced to significantly enhance the edge clarity of the knee joint; the remaining area except the low-sharpening area and the high-sharpening area is a medium-sharpening area, and a medium sharpening coefficient needs to be set to balance the detail retention effect.

[0085] For example, in an embodiment of the present invention, different coefficients are used in the Laplace operator to perform sharpening enhancement on the cartilage collagen fiber area and other areas in the target knee joint MR image, including: using the cartilage collagen fiber area as a low-sharpening mask, recording the area after removing the low-sharpening mask in the edge map of the knee joint MR image as a high-sharpening mask, and using the remaining area except the low-sharpening mask and the high-sharpening mask as a medium-sharpening mask; using the low-sharpening coefficient, the medium-sharpening coefficient and the high-sharpening coefficient in the Laplace operator respectively to perform sharpening enhancement on the low-sharpening mask, the medium-sharpening mask and the high-sharpening mask.

[0086] The Low sharpening factor is smaller than the Medium sharpening factor, and the Medium sharpening factor is smaller than the High sharpening factor. The High sharpening mask and the Medium sharpening mask together form the Other area.

[0087] Specifically, the edge map of the knee joint MR image can be obtained through Canny edge detection.

[0088] For example, the low sharpening factor may range from 0.1 to 0.4; the medium sharpening factor may range from 0.5 to 0.6; and the high sharpening factor may range from 0.7 to 1.0.

[0089] For example, in an embodiment of the present invention, a standard Laplacian convolution kernel may be used to perform convolution on the target knee joint MR image after noise removal to generate a Laplacian response map.

[0090] For example, an embodiment of the present invention provides a formula for calculating a sharpening result. For details, see the following relationship: ; represents the sharpening result of the target knee joint MR image, represents the target knee joint MR image, represents the Laplace response plot, Indicates the sharpening factor.

[0091] The sharpening factor can be a low sharpening factor, a medium sharpening factor, or a high sharpening factor.

[0092] Thus, the embodiment of the present invention has accurately sharpened and enhanced the denoised target knee joint MR image.

[0093] It can be seen that in the embodiment of the present invention, when the knee joint MR image is enhanced, the gradient direction of each pixel point in the knee joint MR image at different scales can be obtained; the degree to which the pixel point belongs to noise is obtained according to the similarity of the gradient direction of the pixel point at different scales, the difference between the amplitude value of the pixel point and the maximum amplitude value in the knee joint MR image; a noise map is generated in response to the comparison result of the degree to which the pixel point belongs to noise and the noise threshold, and the noise map is separated from the knee joint MR image to obtain a target knee joint MR image; the target knee joint MR image is divided into multiple super pixel blocks, and the area of each super pixel block is determined; according to the difference between the area of the super pixel block and the noise threshold, a noise map is generated, and the noise map is separated from the knee joint MR image to obtain a target knee joint MR image; the target knee joint MR image is divided into multiple super pixel blocks, and the area of each super pixel block is determined; The probability that the superpixel block belongs to the cartilage collagen fiber is obtained by comparing the ratio of the minimum circumscribed rectangular area of the superpixel block and the grayscale mean and maximum grayscale values of the superpixel block; in response to the comparison result of the probability that the superpixel block belongs to the cartilage collagen fiber and the probability threshold, a candidate cartilage collagen fiber region is obtained, and the cartilage collagen fiber region is obtained according to the grayscale mean difference between the candidate cartilage collagen fiber region and its surrounding pixels; different coefficients are used in the Laplace operator to perform sharpening enhancement on the cartilage collagen fiber region and other regions in the target knee joint MR image to obtain the enhanced knee joint MR image, which effectively improves the accuracy of knee joint MR image enhancement.

[0094] An embodiment of the present invention further discloses a knee joint MR image enhancement system, comprising a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, a knee joint MR image enhancement method provided by the present invention is implemented.

[0095] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0096] In the present invention, the aforementioned memory may be any tangible medium that contains or stores a program, which may be used by or in combination with an instruction execution system, apparatus, or device.

[0097] The above are all preferred embodiments of the present invention, and are not intended to limit the scope of protection of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A knee joint MR image enhancement method, characterized in that: include: Obtain the gradient direction of each pixel in the knee joint MR image at different scales; The degree to which a pixel belongs to noise is obtained based on the similarity of the gradient direction of the pixel at different scales and the difference between the amplitude value of the pixel and the maximum amplitude value in the knee joint MR image. generating a noise map in response to a comparison result of the degree to which a pixel point belongs to noise and a noise threshold, separating the noise map from the knee joint MR image, and obtaining a target knee joint MR image; Divide the target knee joint MR image into multiple superpixel blocks and determine the area of each superpixel block; The probability that the superpixel block belongs to cartilage collagen fiber is obtained based on the ratio of the area of the superpixel block to the area of the minimum circumscribed rectangle of the superpixel block and the difference between the grayscale mean value and the maximum grayscale value of the superpixel block; In response to a comparison result of the probability that the superpixel block belongs to the cartilage collagen fiber and the probability threshold, a candidate cartilage collagen fiber region is obtained, and the cartilage collagen fiber region is obtained according to the grayscale mean difference between the candidate cartilage collagen fiber region and its surrounding pixels; Different coefficients are used in the Laplace operator to perform sharpening enhancement on the cartilage collagen fiber area and other areas in the target knee joint MR image to obtain an enhanced knee joint MR image.

2. A knee joint MR image enhancement method according to claim 1, characterized in that: The step of obtaining the gradient direction of each pixel in the knee joint MR image at different scales also includes: The knee joint MR image is converted into a spectrum diagram through Fourier transform to obtain the amplitude value of each pixel point and the maximum amplitude value in the knee joint MR image.

3. The knee joint MR image enhancement method according to claim 1, characterized in that: The step of obtaining the gradient direction of each pixel in the knee joint MR image at different scales includes: Different scales are used in the Sobel operator to obtain the gradient direction of each pixel at different scales.

4. The knee joint MR image enhancement method according to claim 1, characterized in that: The degree to which a pixel point belongs to noise is obtained based on the gradient direction similarity between pixels at different scales and the difference between the amplitude value of the pixel point and the maximum amplitude value in the knee joint MR image, including: The mean of the cosine similarity between the gradient direction of each pixel in one scale and the gradient direction of the next scale is recorded as the similarity index of the pixel; the degree to which the i-th pixel belongs to noise is calculated : ； 、 are the amplitude value and similarity index of the i-th pixel, respectively. is the maximum amplitude value in the knee joint MR image, is an exponential function with base e.

5. The knee joint MR image enhancement method according to claim 1, characterized in that: The generating of a noise map in response to a comparison result of the degree to which a pixel point belongs to noise and a noise threshold value comprises: Pixels in the knee joint MR image whose noise level is greater than a noise threshold are regarded as noise points; and a noise map is generated according to the noise points in the knee joint MR image.

6. The knee joint MR image enhancement method according to claim 1, characterized in that: The step of dividing the target knee joint MR image into a plurality of superpixel blocks and determining the area of each superpixel block includes: Determine the seed point in the target knee joint MR image; perform region growing with the seed point as the starting point, and finally divide the target knee joint MR image into multiple superpixel blocks; and use the number of pixels in the superpixel block as the area of the superpixel block.

7. The knee joint MR image enhancement method according to claim 1, characterized in that: The method of obtaining the probability that the superpixel block belongs to cartilage collagen fibers based on the ratio of the area of the superpixel block to the area of the minimum circumscribed rectangle of the superpixel block and the difference between the grayscale mean value and the maximum grayscale value of the superpixel block includes: The super pixel blocks whose gray level mean value in the super pixel block and the maximum gray level mean value of all super pixel blocks are not greater than the difference threshold are screened out; The probability that a superpixel belongs to cartilage collagen fiber : ； For the The grayscale mean of superpixel blocks, is the maximum grayscale value, 、 Respectively The area of the superpixel block, the area of the minimum bounding rectangle, is an exponential function with base e.

8. The knee joint MR image enhancement method according to claim 1, characterized in that: The method of obtaining the cartilage collagen fiber region according to the grayscale mean difference between the candidate cartilage collagen fiber region and its surrounding pixel points includes: The normalized value of the grayscale mean difference between the candidate cartilage collagen fiber region and its surrounding pixel points is recorded as the grayscale contrast of the candidate cartilage collagen fiber region; if the grayscale contrast of any candidate cartilage collagen fiber region is not less than 0, then the grayscale contrasts of the candidate cartilage collagen fiber regions with grayscale contrast not less than 0 are sorted, and the candidate cartilage collagen fiber region with grayscale contrast greater than or equal to the third quartile is selected as the cartilage collagen fiber region.

9. The knee joint MR image enhancement method according to claim 1, characterized in that: The method of using different coefficients to perform sharpening enhancement on the cartilage collagen fiber region and other regions in the target knee joint MR image in the Laplacian operator includes: The cartilage collagen fiber area is used as a low-sharpening mask, the area after removing the low-sharpening mask in the edge map of the knee joint MR image is recorded as a high-sharpening mask, and the remaining area except the low-sharpening mask and the high-sharpening mask is used as a medium-sharpening mask; the low-sharpening coefficient, medium-sharpening coefficient and high-sharpening coefficient are used in the Laplacian operator to perform sharpening enhancement on the low-sharpening mask, medium-sharpening coefficient and high-sharpening mask, respectively; among them, the low-sharpening coefficient is smaller than the medium-sharpening coefficient, and the medium-sharpening coefficient is smaller than the high-sharpening coefficient.

10. A knee joint MR image enhancement system, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a knee joint MR image enhancement method according to any one of claims 1 to 9 is implemented.

Images