Alzheimer's disease medical image processing method and system

Abstract

Disclosed in the present invention are an Alzheimer's disease medical image processing method and system. The method comprises the following steps: calculating statistical features of noise in an Alzheimer's disease medical image to obtain the noise type and distribution; performing denoising processing on the Alzheimer's disease medical image on the basis of the noise type and distribution to obtain a denoised image; performing morphological opening and closing operations on the denoised image to obtain a feature vector of a lesion region; acquiring the lesion region by means of an adaptive threshold segmentation method on the basis of the feature vector of the lesion region; and extracting texture features of the lesion region at different frequencies and directions by means of a multi-scale texture analysis method to obtain texture descriptors. In the present invention, the lesion region and features thereof in the Alzheimer's disease medical image are effectively extracted, providing an important basis for disease diagnosis and analysis.

Description

A method and system for medical image processing in Alzheimer's disease Technical Field

[0001] This invention belongs to the field of image processing technology, and in particular relates to a medical image processing method and system for Alzheimer's disease. Background Technology

[0002] In Alzheimer's disease medical images, lesions often exhibit irregular shapes and blurred boundaries, making them difficult to distinguish from normal tissue areas. This necessitates denoising methods that can adapt to the morphological characteristics of the lesion region, removing noise while preserving as much of the original shape and boundary information as possible, providing a reliable foundation for subsequent segmentation and recognition. However, traditional filtering methods struggle to accommodate these complex morphological features, easily introducing new artifacts and distortions. Therefore, designing an adaptive morphological filtering method that dynamically adjusts filtering parameters based on the local morphological features of the image presents a challenging technical problem.

[0003] Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a medical image processing method and system for Alzheimer's disease, thereby resolving the issues present in the prior art.

[0005] To achieve the above objectives, the present invention provides a medical image processing method for Alzheimer's disease, comprising the following steps:

[0006] Calculate the statistical characteristics of noise in medical images of Alzheimer's disease to obtain the noise type and distribution;

[0007] Denoising is performed on medical images of Alzheimer's disease based on noise type and distribution to obtain denoised images;

[0008] Morphological opening and closing operations are performed on the denoised image to obtain the feature vector of the lesion region;

[0009] Based on the feature vector of the lesion region, the lesion region is obtained through an adaptive threshold segmentation method.

[0010] Texture descriptors are obtained by extracting texture features of lesion regions at different frequencies and directions using multi-scale texture analysis.

[0011] The present invention also provides a medical image processing system for Alzheimer's disease, comprising:

[0012] The noise analysis module is used to calculate the statistical characteristics of noise in medical images of Alzheimer's disease, and to obtain the noise type and distribution.

[0013] An adaptive morphological filtering module is used to denoise medical images of Alzheimer's disease based on the type and distribution of noise, and obtain denoised images.

[0014] The multi-scale morphological feature extraction module is used to perform morphological opening and closing operations on the denoised image to obtain the feature vector of the lesion region.

[0015] The adaptive threshold segmentation module is used to obtain the lesion region based on the feature vector of the lesion region using an adaptive threshold segmentation method.

[0016] The multi-scale texture analysis module is used to extract texture features of the lesion region at different frequencies and directions through multi-scale texture analysis methods to obtain texture descriptors.

[0017] Compared with the prior art, the present invention has the following advantages and technical effects:

[0018] This invention effectively extracts lesion regions and their features from medical images of Alzheimer's disease by combining adaptive filtering, multi-scale morphological operations, and texture analysis, providing important evidence for disease diagnosis and analysis. This method can accurately identify lesions, improving the efficiency and accuracy of Alzheimer's disease diagnosis. Attached Figure Description

[0019] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0020] Figure 1 is a flowchart of the method according to an embodiment of the present invention. Detailed Implementation

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0023] Example 1

[0024] As shown in Figure 1, this embodiment provides a medical image processing method for Alzheimer's disease, including the following steps:

[0025] Calculate the statistical characteristics of noise in medical images of Alzheimer's disease to obtain the noise type and distribution;

[0026] Denoising is performed on medical images of Alzheimer's disease based on noise type and distribution to obtain denoised images;

[0027] Morphological opening and closing operations are performed on the denoised image to obtain the feature vector of the lesion region;

[0028] Based on the feature vector of the lesion region, the lesion region is obtained through an adaptive threshold segmentation method.

[0029] Texture descriptors are obtained by extracting texture features of the lesion region at different frequencies and directions using a multi-scale texture analysis method.

[0030] Furthermore, methods for calculating the statistical characteristics of noise in Alzheimer's disease medical images include:

[0031] Medical images of Alzheimer's disease are preprocessed by normalization and standardization to obtain preprocessed images. Statistical features of the preprocessed images are then extracted, including mean, variance, skewness, and kurtosis.

[0032] Furthermore, methods for obtaining the denoised image include:

[0033] The noise type of the medical image for Alzheimer's disease is determined based on statistical characteristics, and the noise distribution parameters are estimated. The noise distribution parameters are compared with a preset threshold. When the noise exceeds the normal range, the corresponding image denoising algorithm is selected according to the noise type to denoise the medical image for Alzheimer's disease.

[0034] Furthermore, methods for obtaining the feature vector of the lesion region include:

[0035] A first image is obtained by performing morphological opening operations on the denoised image using multi-scale and multi-directional structuring elements; a second image is obtained by performing morphological closing operations on the first image; and a feature vector is constructed based on the second image.

[0036] Furthermore, methods for obtaining the lesion area include:

[0037] Based on the feature vector of the lesion region, an adaptive threshold segmentation algorithm is used to adaptively determine the threshold according to the gray-level histogram of the denoised image. The image to be segmented is binarized, and the lesion region and normal tissue region in the binary image are marked through connected component analysis. The geometric features of the lesion region and normal tissue region are calculated, and the lesion region and normal tissue region are separated according to the feature differences of the geometric features.

[0038] The present invention also provides a medical image processing system for Alzheimer's disease, comprising:

[0039] The noise analysis module is used to calculate the statistical characteristics of noise in medical images of Alzheimer's disease, and to obtain the noise type and distribution.

[0040] An adaptive morphological filtering module is used to denoise medical images of Alzheimer's disease based on the type and distribution of noise, and obtain denoised images.

[0041] The multi-scale morphological feature extraction module is used to perform morphological opening and closing operations on the denoised image to obtain the feature vector of the lesion region.

[0042] The adaptive threshold segmentation module is used to obtain the lesion region based on the feature vector of the lesion region using an adaptive threshold segmentation method.

[0043] The multi-scale texture analysis module is used to extract texture features of the lesion region at different frequencies and directions through multi-scale texture analysis methods to obtain texture descriptors.

[0044] Furthermore, the noise analysis module includes:

[0045] The preprocessing unit is used to perform normalization and standardization preprocessing on Alzheimer's disease medical images to obtain preprocessed images;

[0046] The extraction unit is used to extract statistical features of the preprocessed image, including mean, variance, skewness, and kurtosis.

[0047] Furthermore, the adaptive morphological filtering module includes:

[0048] The noise type acquisition unit is used to determine the noise type of Alzheimer's disease medical images based on statistical characteristics and to estimate the noise distribution parameters.

[0049] The denoising unit compares the noise distribution parameters with a preset threshold. When the noise exceeds the normal range, it selects the appropriate image denoising algorithm according to the noise type to denoise the medical image of Alzheimer's disease.

[0050] Furthermore, the multi-scale morphological feature extraction module includes:

[0051] The opening unit is used to perform morphological opening operations on the denoised image using multi-scale and multi-directional structuring elements to obtain the first image.

[0052] A closing operation unit is used to perform morphological closing operations on the first image to obtain the second image;

[0053] The building unit is used to construct a feature vector based on the second image.

[0054] Furthermore, the adaptive threshold segmentation module includes:

[0055] The threshold determination unit is used to adaptively determine the threshold based on the feature vector of the lesion region and the gray-level histogram of the denoised image using an adaptive threshold segmentation algorithm.

[0056] The segmentation unit is used to binarize the image to be segmented. Through connected component analysis, it marks the lesion region and normal tissue region in the binary image, calculates the geometric features of the lesion region and normal tissue region, and separates the lesion region and normal tissue region according to the feature differences of the geometric features.

[0057] The specific implementation method is as follows:

[0058] First, head MRI or PET image datasets are obtained from a medical image database of Alzheimer's patients. The images are normalized by scaling pixel values ​​to the range of 0-1 and then Z-score standardized to achieve a mean of 0 and a variance of 1. Next, image processing libraries such as OpenCV are used to extract statistical features such as mean, variance, skewness, and kurtosis. For example, an image might have a mean of 52, a variance of 0.8, a skewness of -31, and a kurtosis of 74. These features are then input into a noise type discrimination model based on support vector machines, trained on a large number of labeled noisy images. If the discrimination result is Gaussian noise, a wavelet transform-based Gaussian noise parameter estimation algorithm is used to estimate the mean and variance of the noise. For example, the estimated mean of Gaussian noise might be 0.2 and the variance 0.05. The estimated parameters are then compared with thresholds specified in medical image quality standards. For example, the Gaussian noise variance threshold is 0.03; exceeding this threshold indicates that the noise is outside the normal range. For different types of noise, appropriate denoising algorithms are selected. For example, a nonlocal mean filtering algorithm is used for Gaussian noise to enhance the image. Finally, the denoised image is input into an Alzheimer's disease diagnostic model based on a convolutional neural network. Combined with clinical indicators such as the patient's cognitive function score and cerebrospinal fluid biomarker levels, a logistic regression model is used for comprehensive judgment to obtain the probability of the patient having the disease. If the probability is greater than 6, the patient is diagnosed with Alzheimer's disease. This process, combining medical image analysis with machine learning algorithms, can assist doctors in the early diagnosis of Alzheimer's disease.

[0059] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A medical image processing method for Alzheimer's disease, characterized in that, Includes the following steps: Calculate the statistical characteristics of noise in medical images of Alzheimer's disease to obtain the noise type and distribution; The Alzheimer's disease medical images are denoised according to the noise type and distribution to obtain denoised images. Morphological opening and closing operations are performed on the denoised image to obtain the feature vector of the lesion region; Based on the feature vector of the lesion region, the lesion region is obtained through an adaptive threshold segmentation method; Texture descriptors are obtained by extracting texture features of the lesion region at different frequencies and directions using a multi-scale texture analysis method.

2. The Alzheimer's disease medical image processing method according to claim 1, characterized in that, The method for calculating the statistical characteristics of noise in Alzheimer's disease medical images includes: The medical images of Alzheimer's disease are preprocessed by normalization and standardization to obtain preprocessed images. Statistical features of the preprocessed images are extracted, including mean, variance, skewness, and kurtosis.

3. The Alzheimer's disease medical image processing method according to claim 1, characterized in that, The method for obtaining the denoised image includes: The noise type of the Alzheimer's disease medical image is determined based on the statistical characteristics, and the noise distribution parameters are estimated. The noise distribution parameters are compared with a preset threshold. When the noise exceeds the normal range, the corresponding image denoising algorithm is selected according to the noise type to denoise the Alzheimer's disease medical image.

4. The Alzheimer's disease medical image processing method according to claim 1, characterized in that, The method for obtaining the feature vector of the lesion region includes: A first image is obtained by performing morphological opening operations on the denoised image using multi-scale and multi-directional structuring elements; a second image is obtained by performing morphological closing operations on the first image; and the second image is obtained according to the first... Two images are used to construct feature vectors.

5. The Alzheimer's disease medical image processing method according to claim 1, characterized in that, The method for obtaining the lesion area includes: Based on the feature vector of the lesion region, an adaptive threshold segmentation algorithm is used to adaptively determine the threshold according to the grayscale histogram of the denoised image; the image to be segmented is binarized, and the lesion region and normal tissue region in the binary image are marked through connected component analysis; the geometric features of the lesion region and normal tissue region are calculated; and the lesion region and normal tissue region are separated according to the feature differences of the geometric features.

6. A medical image processing system for Alzheimer's disease, characterized in that, include: The noise analysis module is used to calculate the statistical characteristics of noise in medical images of Alzheimer's disease, and to obtain the noise type and distribution. An adaptive morphological filtering module is used to denoise the Alzheimer's disease medical image according to the noise type and distribution to obtain a denoised image. A multi-scale morphological feature extraction module is used to perform morphological opening and closing operations on the denoised image to obtain the feature vector of the lesion region. An adaptive threshold segmentation module is used to obtain the lesion region based on the feature vector of the lesion region using an adaptive threshold segmentation method. The multi-scale texture analysis module is used to extract texture features of the lesion region at different frequencies and directions through multi-scale texture analysis methods to obtain texture descriptors.

7. The Alzheimer's disease medical image processing system according to claim 6, characterized in that, The noise analysis module includes: The preprocessing unit is used to perform normalization and standardization preprocessing on the Alzheimer's disease medical images to obtain preprocessed images; The extraction unit is used to extract statistical features of the preprocessed image, including mean, variance, skewness, and kurtosis.

8. The Alzheimer's disease medical image processing system according to claim 6, characterized in that, The adaptive morphological filtering module includes: The noise type acquisition unit is used to determine the noise type of the Alzheimer's disease medical image based on the statistical characteristics and to estimate the noise distribution parameters. The denoising unit is used to compare the noise distribution parameters with a preset threshold. When the noise exceeds the normal range, the corresponding image denoising algorithm is selected according to the noise type to denoise the medical image of Alzheimer's disease.

9. The Alzheimer's disease medical image processing system according to claim 6, characterized in that, The multi-scale morphological feature extraction module includes: An opening unit is used to perform morphological opening operations on the denoised image using multi-scale and multi-directional structuring elements to obtain a first image; A closing operation unit is used to perform morphological closing operations on the first image to obtain a second image; A construction unit is used to construct a feature vector based on the second image.

10. The Alzheimer's disease medical image processing system according to claim 6, characterized in that, The adaptive threshold segmentation module includes: The threshold determination unit is used to adaptively determine the threshold based on the feature vector of the lesion region and the gray-level histogram of the denoised image using an adaptive threshold segmentation algorithm. The segmentation unit is used to binarize the image to be segmented, mark the lesion region and normal tissue region in the binary image through connected component analysis, calculate the geometric features of the lesion region and normal tissue region, and separate the lesion region and normal tissue region according to the feature differences of the geometric features.

Images