Medical image segmentation method and system based on fuzzy C-means clustering, and medium

Abstract

The invention discloses a medical image segmentation method and system based on fuzzy C-means clustering, and a medium, and belongs to the technical field of medical image segmentation. The method comprises the following steps: an image preprocessing stage: performing brain surface extraction, image noise reduction and image normalization on an input initial brain image to obtain a first image; a pixel clustering stage: performing pixel clustering on the first image by using a fuzzy C-means clustering algorithm, setting clustering parameters, calculating a membership matrix and a cluster center of the first image, and continuously updating and iterating to perform dynamic adjustment until a convergence condition is met to obtain a second image; and in the image segmentation stage, according to the membership degree of each pixel to each cluster in the second image, the cluster to which the pixel belongs is determined, and the cluster with the maximum membership degree of each pixel point is selected as the classification result of the pixel. Through dynamic adjustment of the fuzzy membership function, different tissue types can be separated more accurately under the condition that the image quality is not high or the boundary is fuzzy.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image segmentation, and in particular to a medical image segmentation method, system and medium based on fuzzy C-means clustering. Background Art

[0002] Medical image segmentation is an important research direction in the field of medical image processing. Its goal is to separate the region of interest (ROI) in the image from the background for further analysis, diagnosis and treatment. With the rapid development of medical imaging technology, the acquisition of medical images such as computed tomography (CT), magnetic resonance imaging (MRI), ultrasound and X-ray has become an indispensable tool in clinical diagnosis and surgical intervention. However, these medical images often face multiple challenges in practical applications, making the image segmentation task very complex and highly challenging. Among them, aliasing, mixed pixels, noise interference and blurring are common quality problems in medical images. Aliasing refers to the loss or distortion of image details due to insufficient sampling rate, which affects the accuracy of segmentation results. The problem of mixed pixels stems from the limitation of image resolution, which causes the pixel information of adjacent tissues or structures to merge together and is difficult to separate accurately. Noise interference and blurring are due to physical interference (such as movement, equipment instability, etc.) that may occur during the imaging process, resulting in a decrease in image quality, which in turn affects the accuracy of subsequent processing and analysis. Therefore, how to effectively deal with these problems and extract clear and accurate image information has become a core problem in medical image segmentation.

[0003] As one of the most complex components of the brain, the fine segmentation of its tissue structure is crucial for clinicians to make accurate diagnosis and therapeutic intervention. However, problems such as aliasing, mixed pixels, noise interference and blurring effects are common in medical images, especially in brain magnetic resonance imaging (MRI), because the boundaries between different brain tissues (such as gray matter, white matter and cerebrospinal fluid) are often blurred and difficult to distinguish clearly, making brain MRI segmentation tasks particularly challenging. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and to provide a medical image segmentation method, system and medium based on fuzzy C-means clustering.

[0005] The objective of the present invention is achieved through the following technical solutions: The first aspect of the present invention provides: a medical image segmentation method based on fuzzy C-means clustering, comprising the following steps: In the image preprocessing stage, the input initial brain image is subjected to brain surface extraction, image denoising and image normalization to obtain a first image; In the pixel clustering stage, the fuzzy C-means clustering algorithm is used to cluster the pixels of the first image, set clustering parameters, calculate the membership matrix and cluster centers of the first image, and continuously update and iterate to perform dynamic adjustments until the convergence conditions are met to obtain the second image; In the image segmentation stage, the cluster to which each pixel belongs is determined according to its membership to each cluster in the second image, and the cluster with the largest membership of each pixel is selected as the classification result of the pixel.

[0006] Preferably, the brain surface extraction comprises the following steps: Performing alignment processing on the initial brain image; Use the brain surface extraction tool to extract the outer contour image of the brain, and remove the skull, skin and other non-brain tissue images to obtain the precise boundaries of the brain area; Then the segmentation of the gray matter brain tissue image, the white matter brain tissue image and the cerebrospinal fluid brain tissue image is optimized, and finally the gray matter brain tissue image, the white matter brain tissue image and the cerebrospinal fluid brain tissue image are classified according to their signal intensity and position in the initial brain image.

[0007] Preferably, the image noise reduction comprises the following steps: Use wavelet denoising to remove noise from the input image signal and retain the target features of the input image signal. Suppose the signal destroyed by noise is S(t): S(t)= Y(t)+ N(t), where Y(t) represents the original pixel signal and N(t) represents the noise signal; The approximate original signal is obtained using the following formula : ,in and They represent the forward and inverse operations of the wavelet operator, respectively. represents the denoising operator with threshold 𝛾.

[0008] Preferably, the image normalization comprises the following steps: Each pixel value of the initial brain image P Normalization is performed according to the following formula: ,in P norm is the normalized pixel value, P max is the maximum pixel value, R max is the maximum value of pixel bit depth, P min is the minimum pixel value.

[0009] Preferably, the pixel clustering stage further comprises the following steps: Set the number of clusters to be divided into C to 3, and mark them as gray matter, white matter, and cerebrospinal fluid categories respectively; set the fuzzy factor to control the fuzziness of the membership degree; The membership degree of each pixel is calculated to indicate the degree to which it belongs to each cluster; the distance from each pixel to the center of each cluster is calculated; and the Euclidean distance between each pixel and the center of the initial cluster is calculated based on the pixel value in the first image: ;in, p i is the i-th pixel, c j is the jth cluster center, d is the dimension of the image, p i,k is the value of the i-th pixel in the k-th dimension, c i,k is the value of the j-th cluster center in the k-th dimension; Update the membership value of each pixel to each cluster according to the Euclidean distance: ;in, u i,j Represents the membership of the i-th pixel to the j-th cluster, P i Represents the position vector of the i-th pixel, v j is the position vector representing the center of the jth cluster, v k represents the position vector of the kth cluster center, m Represents the fuzzy index used to control the fuzziness degree of membership; According to the updated membership matrix, the center of each cluster is recalculated. The cluster center is the weighted average of all pixels in each cluster. The calculation formula is: ;in N Represents the total number of pixels. The power of the membership of the i-th pixel to the j-th cluster, where m is the fuzzy index; The cluster center and the membership matrix are continuously updated. When the change of the membership matrix is ​​less than a preset threshold, it is considered that the convergence condition is met, and the iteration is stopped to obtain the second image.

[0010] Preferably, the fuzzy factor is 2.

[0011] Preferably, the membership matrix is ​​retained in the classification result to represent the membership of each pixel point to different brain tissue types so as to achieve fuzzy classification.

[0012] The second aspect of the present invention provides: a medical image segmentation system based on fuzzy C-means clustering, used to implement any of the above-mentioned medical image segmentation methods based on fuzzy C-means clustering, comprising: An image preprocessing module, used for performing brain surface extraction, image noise reduction and image normalization on the input initial brain image to obtain a first image; A pixel clustering module is used to cluster pixels of the first image using a fuzzy C-means clustering algorithm, set clustering parameters, calculate the membership matrix and cluster centers of the first image, and continuously update and iterate to perform dynamic adjustments until a convergence condition is met to obtain a second image; The image segmentation module is used to determine the cluster to which each pixel belongs according to the membership degree of each pixel in the second image to each cluster, and select the cluster with the largest membership degree of each pixel as the classification result of the pixel.

[0013] The third aspect of the present invention provides: a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are loaded and executed by a processor, any of the above-mentioned medical image segmentation methods based on fuzzy C-means clustering is implemented.

[0014] The beneficial effects of the present invention are: 1) Through dynamic adjustment of the fuzzy membership function, different tissue types can be separated more accurately when the image quality is not high or the boundary is blurred.

[0015] 2) By enhancing the processing of image noise, aliasing effects, mixed pixels and other problems, the image quality is improved, so that this method can effectively improve the segmentation effect and ensure the accuracy on high-noise, low-quality or blurred images.

[0016] 3) It can reduce human intervention, lower the risk of misdiagnosis, and provide reliable data support for precision medicine and personalized treatment. It has important application value, especially in the early diagnosis and treatment planning of brain diseases; it provides technical support for precision medicine and intelligent medical image processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of the medical image segmentation method based on fuzzy C-means clustering; Figure 2 Comparison diagram of the initial brain image, the preprocessed brain image, and the segmented brain image. DETAILED DESCRIPTION

[0018] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0019] The present invention aims to solve the problem of fuzzy boundaries and mixed pixels between different tissues (such as gray matter, white matter and cerebrospinal fluid) in magnetic resonance (MR) brain images. By introducing fuzzy membership to enhance the handling of image uncertainty and ambiguity, the method can dynamically adjust the membership function to better distinguish the transition areas of different tissues. Through these improvements, the method shows higher accuracy in the segmentation of brain MR images, especially in the case of poor image quality or blurred boundaries, it can effectively distinguish tissues such as gray matter, white matter and cerebrospinal fluid, and can provide support for personalized treatment and precision medicine.

[0020] The medical image segmentation method based on fuzzy C-means clustering proposed in the present invention has significant advantages, especially when processing brain MR images. First, the method enables the segmentation process to dynamically adjust the membership function, especially when the image boundary is blurred or there are mixed pixels, it can more accurately separate different tissue types (such as gray matter, white matter and cerebrospinal fluid), thereby improving the accuracy and robustness of the segmentation. These improvements enable the method to effectively improve the image segmentation effect when processing high-noise, low-quality or blurred medical images, and ensure the accuracy of image analysis in clinical diagnosis and treatment. Therefore, the application of this technical solution can significantly improve the automation and intelligence level of medical image processing, reduce the need for manual intervention, and thus reduce the labor cost and misdiagnosis risk in clinical imaging diagnosis. At the same time, accurate image segmentation technology can provide more reliable data support for personalized treatment and precision medicine, promote the progress of precision medicine, and thus improve the utilization efficiency of medical resources and the treatment effect. This efficient segmentation method can also be widely used in the early diagnosis and treatment planning of brain diseases, improve the diagnostic ability and treatment level of hospitals, bring greater economic benefits, and help reduce unnecessary expenses in the treatment process and improve the treatment effect of patients.

[0021] See also Figure 1-Figure 2 The first aspect of the present invention provides: a medical image segmentation method based on fuzzy C-means clustering, comprising the following steps: In the image preprocessing stage, the input initial brain image is subjected to brain surface extraction, image denoising and image normalization to obtain a first image; In the pixel clustering stage, the fuzzy C-means clustering algorithm is used to cluster the pixels of the first image, set clustering parameters, calculate the membership matrix and cluster centers of the first image, and continuously update and iterate to perform dynamic adjustments until the convergence conditions are met to obtain the second image; In the image segmentation stage, the cluster to which each pixel belongs is determined according to its membership to each cluster in the second image, and the cluster with the largest membership of each pixel is selected as the classification result of the pixel.

[0022] In this embodiment, first, the brain image is preprocessed to improve the quality of feature extraction and image analysis results; secondly, an improved fuzzy C-means clustering is proposed to group pixels; then, the clusters are assigned according to the distance from the jth cluster to the i-th data object using the improved distance norm; finally, the brain image is finally segmented into gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) tissue.

[0023] Image preprocessing is an indispensable key step to improve image quality and prepare for subsequent analysis. In medical image analysis, preprocessing can not only improve image quality and eliminate artifacts, but also enhance image features, thereby ensuring that the data provides a reliable basis for accurate diagnosis, treatment planning, and scientific research purposes. Before processing the cluster analysis of brain images, the first task is to accurately extract brain regions from MRI images. To ensure the accuracy of the segmentation results, this method uses a series of preprocessing steps to deal with problems such as inhomogeneity and partial volume effects in the image to improve the accuracy of subsequent analysis.

[0024] In some embodiments, the brain surface extraction comprises the following steps: Performing alignment processing on the initial brain image; Use the brain surface extraction tool to extract the outer contour image of the brain, and remove the skull, skin and other non-brain tissue images to obtain the precise boundaries of the brain area; Then the segmentation of the gray matter brain tissue image, the white matter brain tissue image and the cerebrospinal fluid brain tissue image is optimized, and finally the gray matter brain tissue image, the white matter brain tissue image and the cerebrospinal fluid brain tissue image are classified according to their signal intensity and position in the initial brain image.

[0025] In this embodiment, the main components of brain tissue include gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF). However, in addition to brain tissue, brain images also contain many other types of tissues, such as fat, muscle, skin, skull, glial substances and connective tissue. In order to perform accurate brain tissue segmentation, these non-brain tissues must be extracted from the image first. First, the brain image needs to be properly aligned to ensure the spatial consistency of the image and reduce errors caused by differences in scanning position and posture. The aligned image can effectively help the segmentation algorithm to more accurately identify the boundaries and structures between different tissues. Next, for the extraction of non-brain tissue, the method uses some specialized brain surface extraction tools (such as MIPAV) to automatically identify and separate brain tissue and non-brain tissue, further improving the accuracy of segmentation. Specifically, the outer contour of the brain is first extracted by an image processing algorithm, and the precise boundaries of the brain region are obtained by removing the skull, skin and other non-brain tissues. Subsequently, the segmentation of brain tissues such as gray matter, white matter and cerebrospinal fluid is further optimized, and these brain tissues are classified according to their signal intensity and position in the MRI image.

[0026] In some embodiments, the image noise reduction comprises the following steps: Use wavelet denoising to remove noise from the input image signal and retain the target features of the input image signal. Suppose the signal destroyed by noise is S(t): S(t)= Y(t)+ N(t), where Y(t) represents the original pixel signal and N(t) represents the noise signal; The approximate original signal is obtained using the following formula : ,in and They represent the forward and inverse operations of the wavelet operator, respectively. represents the denoising operator with threshold 𝛾.

[0027] In some embodiments, the image normalization comprises the following steps: Each pixel value of the initial brain image P Normalization is performed according to the following formula: ,in P norm is the normalized pixel value, P max is the maximum pixel value, R max is the maximum value of pixel bit depth, P min is the minimum pixel value.

[0028] In some embodiments, the pixel clustering stage further comprises the following steps: Set the number of clusters to be divided into C to 3, and mark them as gray matter, white matter, and cerebrospinal fluid categories respectively; set the fuzzy factor to control the fuzziness of the membership degree; The membership degree of each pixel is calculated to indicate the degree to which it belongs to each cluster; the distance from each pixel to the center of each cluster is calculated; and the Euclidean distance between each pixel and the center of the initial cluster is calculated based on the pixel value in the first image: ;in, p i is the i-th pixel, c j is the jth cluster center, d is the dimension of the image, p i,k is the value of the i-th pixel in the k-th dimension, c i,k is the value of the j-th cluster center in the k-th dimension; Update the membership value of each pixel to each cluster according to the Euclidean distance: ;in, u i,j Represents the membership of the i-th pixel to the j-th cluster, P i Represents the position vector of the i-th pixel, v j is the position vector representing the center of the jth cluster, v k represents the position vector of the kth cluster center, m Represents the fuzzy index used to control the fuzziness degree of membership; According to the updated membership matrix, the center of each cluster is recalculated. The cluster center is the weighted average of all pixels in each cluster. The calculation formula is: ;in N Represents the total number of pixels. The power of the membership of the i-th pixel to the j-th cluster, where m is the fuzzy index; The cluster center and the membership matrix are continuously updated. When the change of the membership matrix is ​​less than a preset threshold, it is considered that the convergence condition is met, and the iteration is stopped to obtain the second image.

[0029] In some embodiments, the fuzziness factor is 2.

[0030] In some embodiments, a membership matrix is ​​retained in the classification result to represent the membership of each pixel to different brain tissue types, thereby achieving fuzzy classification.

[0031] In this embodiment, it is possible to choose whether to retain the membership matrix to achieve hard classification or fuzzy classification. If fuzzy information needs to be retained, the membership matrix can be retained to indicate the membership of each pixel to different brain tissue types. Figure 2 As shown in the figure, the first column is the original image, the second column is the preprocessed image, and the third column is the segmented image. By preprocessing the image, the quality of feature extraction and analysis is improved, and the boundary blur and aliasing that may occur in traditional methods are effectively reduced; further, the brain tissue (gray matter, white matter and cerebrospinal fluid) is accurately classified and segmented through fuzzy C-means clustering.

[0032] The second aspect of the present invention provides: a medical image segmentation system based on fuzzy C-means clustering, used to implement any of the above-mentioned medical image segmentation methods based on fuzzy C-means clustering, comprising: An image preprocessing module, used for performing brain surface extraction, image noise reduction and image normalization on the input initial brain image to obtain a first image; A pixel clustering module is used to cluster pixels of the first image using a fuzzy C-means clustering algorithm, set clustering parameters, calculate the membership matrix and cluster centers of the first image, and continuously update and iterate to perform dynamic adjustments until a convergence condition is met to obtain a second image; The image segmentation module is used to determine the cluster to which each pixel belongs according to the membership degree of each pixel in the second image to each cluster, and select the cluster with the largest membership degree of each pixel as the classification result of the pixel.

[0033] The third aspect of the present invention provides: a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are loaded and executed by a processor, any of the above-mentioned medical image segmentation methods based on fuzzy C-means clustering is implemented.

[0034] The present invention mainly targets the problem of fuzzy boundaries and mixed pixels between different tissues (including gray matter, white matter and cerebrospinal fluid) in brain MIR images. The key point is that by introducing a fuzzy membership function, the processing of image uncertainty and fuzziness can be dynamically adjusted, thereby improving the segmentation accuracy, especially in the case of poor image quality, blurred boundaries or noise interference. This method effectively solves the limitations of traditional segmentation technology in processing low-quality medical images and improves the accuracy and robustness of segmentation results. Especially when processing brain MR images, it can effectively overcome the multiple challenges of traditional image segmentation technology in practical applications. First, by introducing fuzzy membership and dynamically adjusting the membership function, the method can effectively handle the problem of fuzzy boundaries and mixed pixels in the image. Traditional image segmentation methods often assume that pixels belong to a specific category, but in medical images, especially brain images, the boundaries of tissues such as gray matter, white matter and cerebrospinal fluid are often fuzzy, and transition areas and mixed pixels are easily generated, resulting in reduced segmentation accuracy. The fuzzy C-means clustering method can achieve a smooth transition between different tissues by assigning a membership value to pixels, avoiding boundary distortion caused by hard segmentation, and making the separation of various tissues more accurate. Secondly, the method has strong robustness, especially when facing low-quality, high-noise or blurred medical images, it can still maintain a high segmentation accuracy. Medical images are often affected by factors such as equipment instability and patient movement, resulting in image quality degradation or noise, which poses challenges to subsequent image analysis and diagnosis. The fuzzy C-means method can effectively reduce the interference of noise on the segmentation results through a fuzzy membership model, and can still accurately distinguish different tissue types in the case of poor image quality, thereby ensuring the accuracy of diagnosis in clinical applications. In addition, this technical method has also significantly improved the automation and intelligence level of medical image processing. Traditional medical image segmentation often relies on a lot of manual intervention, which is time-consuming and labor-intensive and easily affected by human factors. After adopting the fuzzy C-means clustering method, the segmentation process can be completed automatically with less manual intervention, greatly reducing the complexity and error rate of manual operations, reducing labor costs, and improving work efficiency. This is of great significance for the processing of large-scale medical images, especially the real-time requirements in clinical applications.

[0035] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.

Claims

1. A medical image segmentation method based on fuzzy C-means clustering, characterized by: The following steps are involved: In the image preprocessing stage, the input initial brain image is subjected to brain surface extraction, image denoising and image normalization to obtain a first image; In the pixel clustering stage, the fuzzy C-means clustering algorithm is used to cluster the pixels of the first image, set clustering parameters, calculate the membership matrix and cluster centers of the first image, and continuously update and iterate to perform dynamic adjustments until the convergence conditions are met to obtain the second image; In the image segmentation stage, the cluster to which each pixel belongs is determined according to its membership to each cluster in the second image, and the cluster with the largest membership of each pixel is selected as the classification result of the pixel.

2. The medical image segmentation method based on fuzzy C-means clustering according to claim 1, characterized in that: The brain surface extraction comprises the following steps: Performing alignment processing on the initial brain image; Use the brain surface extraction tool to extract the outer contour image of the brain, and remove the skull, skin and other non-brain tissue images to obtain the precise boundaries of the brain area; Then the segmentation of the gray matter brain tissue image, the white matter brain tissue image and the cerebrospinal fluid brain tissue image is optimized, and finally the gray matter brain tissue image, the white matter brain tissue image and the cerebrospinal fluid brain tissue image are classified according to their signal intensity and position in the initial brain image.

3. The medical image segmentation method based on fuzzy C-means clustering according to claim 1, characterized in that: The image noise reduction comprises the following steps: Use wavelet denoising to remove noise from the input image signal and retain the target features of the input image signal. Suppose the signal destroyed by noise is S(t): S(t)= Y(t)+ N(t), where Y(t) represents the original pixel signal and N(t) represents the noise signal; The approximate original signal is obtained using the following formula : ,in and They represent the forward and inverse operations of the wavelet operator, respectively. represents the denoising operator with threshold 𝛾.

4. The medical image segmentation method based on fuzzy C-means clustering according to claim 1, characterized in that: The image normalization comprises the following steps: Each pixel value of the initial brain image P Normalization is performed according to the following formula: ,in P norm is the normalized pixel value, P max is the maximum pixel value, R max is the maximum value of pixel bit depth, P min is the minimum pixel value.

5. The medical image segmentation method based on fuzzy C-means clustering according to claim 1, characterized in that: The pixel clustering stage further comprises the following steps: Set the number of clusters to be divided into C to 3, and mark them as gray matter, white matter, and cerebrospinal fluid categories respectively; set the fuzzy factor to control the fuzziness of the membership degree; The membership degree of each pixel is calculated to indicate the degree to which it belongs to each cluster; the distance from each pixel to the center of each cluster is calculated; and the Euclidean distance between each pixel and the center of the initial cluster is calculated based on the pixel value in the first image: ;in, p i is the i-th pixel, c j is the jth cluster center, d is the dimension of the image, p i,k is the value of the i-th pixel in the k-th dimension, c i,k is the value of the j-th cluster center in the k-th dimension; Update the membership value of each pixel to each cluster according to the Euclidean distance: ;in, u i,j Represents the membership of the i-th pixel to the j-th cluster, P i Represents the position vector of the i-th pixel, v j is the position vector representing the center of the jth cluster, v k represents the position vector of the kth cluster center, m Represents the fuzzy index used to control the fuzziness degree of membership; According to the updated membership matrix, the center of each cluster is recalculated. The cluster center is the weighted average of all pixels in each cluster. The calculation formula is: ;in N Represents the total number of pixels. The power of the membership of the i-th pixel to the j-th cluster, where m is the fuzzy index; The cluster center and the membership matrix are continuously updated. When the change of the membership matrix is ​​less than a preset threshold, it is considered that the convergence condition is met, and the iteration is stopped to obtain the second image.

6. The medical image segmentation method based on fuzzy C-means clustering according to claim 5, characterized in that: The fuzzy factor is 2.

7. The medical image segmentation method based on fuzzy C-means clustering according to claims 1-6, characterized in that: The membership matrix is ​​retained in the classification results to represent the membership of each pixel to different brain tissue types, thereby achieving fuzzy classification.

8. A medical image segmentation system based on fuzzy C-means clustering, characterized by: The method for medical image segmentation based on fuzzy C-means clustering according to any one of claims 1 to 7 comprises: An image preprocessing module, used for performing brain surface extraction, image noise reduction and image normalization on the input initial brain image to obtain a first image; A pixel clustering module is used to cluster pixels of the first image using a fuzzy C-means clustering algorithm, set clustering parameters, calculate the membership matrix and cluster centers of the first image, and continuously update and iterate to perform dynamic adjustments until a convergence condition is met to obtain a second image; The image segmentation module is used to determine the cluster to which each pixel belongs according to the membership degree of each pixel in the second image to each cluster, and select the cluster with the largest membership degree of each pixel as the classification result of the pixel.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are loaded and executed by the processor, the medical image segmentation method based on fuzzy C-means clustering as described in any one of claims 1 to 7 is implemented.

Images