Optimized automatic CT image enhancement method based on Windows and histogram equalization

Through the method based on Windowing and histogram equalization, consistency, bone occlusion, low contrast and noise in CT images are solved, and the optimization and enhancement of CT images are achieved, improving the visibility of anatomical structure and diagnostic accuracy.

CN118918014APending Publication Date: 2024-11-08ZHUHAI HENGQIN ALL-STAR MEDICAL TECHNOLOGY CO LTD
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202410776840.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-17
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

There are problems such as difficulty in visualizing anatomical structures in CT images, skeletal structure blocking soft tissue, low contrast, poor window width and window position settings, inaccurate image feature segmentation, and noise interference, which affects the diagnostic accuracy.

Method used

Using optimized automatic CT image enhancement methods based on Windowing and histogram equalization, including standardized processing, suppression of bone structure, contrast enhancement, window width and window adjustment, building hybrid Gaussian models, obtaining standardized gradient histograms and Gaussian smoothing, and maximum Intensity projection visualization.

Benefits of technology

Improve the visibility and diagnostic accuracy of anatomical structures, reduce the impact of bone artifacts, enhance contrast, smooth noise, highlight specific tissue areas, improve visibility of lesion areas and treatment monitoring effects.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118918014A_ABST
    Figure CN118918014A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of medical imaging, in particular to an optimized automatic CT image enhancement method based on Windows and histogram equalization, and the method comprises the following specific steps: standardization, skeleton inhibition, contrast enhancement by using CLAHE, window width and window level adjustment, construction of a Gaussian mixture model, obtaining of a standardized gradient histogram, and smoothing. And the maximum intensity projection is used for visualization. The processed CT image has the following effects: the pixel value of the input CT image is standardized; the influence of skeleton artifacts on the CT image is reduced; enhancing the local contrast of the CT image; a specific interested tissue region can be highlighted by adjusting and changing the display intensity range of the CT image; the contrast and the brightness of the CT image are improved; a Gaussian mixture model is obtained, and gradient features of different areas in the CT image are accurately captured; a standard gradient histogram is obtained, so that the edge features of the CT image are more consistent; the noise of the CT image is further reduced and the definition is enhanced; and visualizing the segmented or enhanced CT image.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of medical imaging, and in particular to an optimized automatic CT image enhancement method based on Windowing and histogram equalization. Background Art

[0002] Computed tomography (CT) imaging can provide detailed cross-sectional images of the human body and is an important diagnostic tool widely used in the medical field. CT images used for diagnosis usually need to be enhanced to improve the visualization of anatomical structures, so as to improve the accuracy of lesion area identification and diagnosis, and provide an important basis for personalized treatment. However, the visualization of anatomical structures in CT images still faces many challenges, as follows:

[0003] 1. Inconsistent intensity values ​​between different CT scans make image consistency comparison and analysis difficult.

[0004] 2. High-strength bone structures may obscure soft tissues, making identification and analysis of these areas difficult.

[0005] 3. Soft tissues usually have low contrast in CT images, making it difficult to distinguish different anatomical features.

[0006] 4. Standard CT images may have poor visualization of certain anatomical features due to poor window width (WW) and window level (WL) settings.

[0007] 5. Due to the complexity of CT images, the segmentation of image features (such as edges and boundaries) may not be accurate.

[0008] 6. Noise and image artifacts may obscure important details in CT images, reducing the clarity and accuracy of visualization.

[0009] To address these challenges, an optimized automatic CT image enhancement method based on windowing and histogram equalization was developed. The method includes multiple key steps aimed at normalizing, suppressing irrelevant structures, enhancing contrast, and more effectively segmenting image features. Summary of the invention

[0010] In response to the challenges mentioned in the background technology section, the present invention provides an innovative method, an optimized automatic CT image enhancement method based on Windowing and histogram equalization, for enhancing CT images.

[0011] The invention discloses an optimized automatic CT image enhancement method based on Windowing and histogram equalization, which adopts the following technical scheme: Standardization: using the minimum-maximum normalization formula to normalize the intensity value of each pixel of the CT image to the range of [0,1];

[0012] Suppress bones: suppress pixels with high intensity bone structures by applying a threshold;

[0013] Contrast enhancement using CLAHE: Apply Contrast-Limited Adaptive Histogram Equalization (CLAHE);

[0014] Window width and window level adjustment: adjust the window width (WW) and window level (WL) settings;

[0015] Construct a mixed Gaussian model: Construct a mixed Gaussian model (GMM) based on the gradient distribution;

[0016] Get the normalized gradient histogram;

[0017] Smoothing: Apply Gaussian smoothing;

[0018] Visualization using Maximum Intensity Projection: The enhanced or segmented images are visualized using the Maximum Intensity Projection (MIP) technique; together, these steps constitute an optimized automatic CT image enhancement method based on windowing and histogram equalization, which improves the visibility of key anatomical structures and facilitates accurate diagnosis and analysis.

[0019] The standardized method is:

[0020] Minimum-maximum normalization formula:

[0021]

[0022] In the formula, I norm (x, y) is the normalized intensity value after standardization, I(x, y) is the original intensity value of the pixel with coordinates (x, y) in the CT image, and I min is the minimum intensity value in the image, I max is the maximum intensity value in the image.

[0023] The method of suppressing pixels of high-intensity bone structures by applying a threshold is:

[0024] Inhibition formula:

[0025]

[0026] T=Threshold for bome suppressiom

[0027] In the formula, I suppressed (x, y) is the intensity value after suppression, I(x, y) is the original intensity value of the pixel with coordinates (x, y) in the CT image, and T is the threshold for distinguishing between bones (high intensity) and soft tissues (low intensity).

[0028] The method of applying Contrast Limited Adaptive Histogram Equalization (CLAHE) is:

[0029] Contrast Enhancement using CLAHE formula:

[0030]

[0031] In the formula, I is the original image, I i,j To divide the original image into several non-overlapping sub-images, i, j represent the row and column indexes of the sub-image, m×n is set to the size of the original image, and a×b is set to the size of each sub-image.

[0032]

[0033] In the formula, H i,j (k) is sub-graph I i,j Grayscale histogram, k represents the grayscale level (0-255), δ is the Kroneckerdelta function, when I(x,y)=k, δ=1, otherwise δ=0.

[0034]

[0035] In the formula, clipLimit is the set contrast limit threshold, and the part exceeding clipLimit is evenly distributed to other gray levels.

[0036]

[0037] In the formula, is the enhanced grayscale value, obtained by looking up the table.

[0038]

[0039] In the formula, I enhanced is the enhanced original image combined from all sub-images, For sub-graph.

[0040] The method for adjusting the window width and window level is:

[0041] Set the upper and lower limits of the window width and window level to remap the intensity value of each pixel. Pixel values ​​below the lower limit are displayed in black, those above the upper limit are displayed in white, and pixel values ​​between the upper and lower limits are linearly mapped to the grayscale range.

[0042] The method for constructing the mixed Gaussian model is:

[0043]

[0044] Where G is the gradient size of the pixel and θ is the direction of each pixel.

[0045] Prepare the gradient data of each pixel (G(x,y),θ(x,y)) to be input into GMM (Gaussian Mixture Model); initialize GMM and use the expectation maximization (EM) algorithm to iteratively optimize the parameters of GMM;

[0046] Expectation-Maximization (EM) algorithm:

[0047] E step (expected step):

[0048]

[0049] In the formula, π k is the mixing coefficient of the kth Gaussian component, Ν(x i μ k ,∑ k ) is the probability density function of the kth Gaussian component.

[0050] M step (maximization step):

[0051]

[0052] In the formula, μ k is the mean of the kth Gaussian component, Σ k is the covariance matrix, γ ik is the probability that the kth data point belongs to the kth component, and N is the total number of data points.

[0053] The method for obtaining the standardized gradient histogram is:

[0054] The Sobel operator is used to calculate the gradient value of each pixel in the image, and then the gradient strength and direction of each pixel are calculated. The calculated gradient values ​​are then standardized and assigned to different gradient intervals to form a gradient histogram, which is then smoothed.

[0055] The smoothing method:

[0056] Gaussian smoothing formula:

[0057]

[0058] Where, Gaussian kernel G(x,y) is a two-dimensional normal distribution function, σ is the standard deviation, and (x,y) is the coordinate of the kernel.

[0059]

[0060] Where i, j are the indices of the Gaussian kernel matrix, and k is the center index of the kernel (i.e. ).

[0061]

[0062] In the formula, I smooth (x,y) is the smoothed image, G(i,j) is the weight of the Gaussian kernel, I(xi,yj) is the pixel value of the original image, and k is the size of the kernel.

[0063] The method described for visualization using maximum intensity projection:

[0064] Select the projection direction and use MIP (Maximum Intensity Projection) for the data source to obtain the processed image. When displaying the image, you can adjust the brightness and contrast or apply color mapping to obtain better visual effects.

[0065] The beneficial effects of the technical solution of the present invention are: obtaining a CT image with standardized pixel values ​​through standardization processing, eliminating the influence of image acquisition settings or equipment differences; reducing the influence of bone artifacts in CT images by suppressing bone structures; enhancing the local contrast of CT images through CLAHE without amplifying noise; obtaining a CT image that can highlight a specific tissue area of ​​interest by changing the image display intensity range through window width and window position adjustment; optimizing the parameters of the mixed Gaussian model through the expectation maximization algorithm to construct a mixed Gaussian model; obtaining a standardized gradient histogram to make the edge features of the CT image more consistent; further reducing CT image noise through Gaussian smoothing; and visualizing the segmented or enhanced CT image through maximum intensity projection. To summarize the above effects, the practical significance of the present invention is: improving the visibility of the lesion area, improving the diagnostic accuracy and achieving better treatment monitoring effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0067] Figure 1 It is a flow chart of an optimized automatic CT image enhancement method based on Windowing and histogram equalization of the present invention;

[0068] Figure 2 is a CT image used as an input source in an optimized automatic CT image enhancement method based on windowing and histogram equalization of the present invention;

[0069] Figure 3It is a CT image that has been standardized in an optimized automatic CT image enhancement method based on windowing and histogram equalization of the present invention;

[0070] Figure 4 It is a CT image processed by suppressing bones in an optimized automatic CT image enhancement method based on windowing and histogram equalization of the present invention;

[0071] Figure 5 A CT image after contrast enhancement using CLAHE in an optimized automatic CT image enhancement method based on Windowing and histogram equalization of the present invention;

[0072] Figure 6 It is a CT image processed by adjusting the window width and window position in an optimized automatic CT image enhancement method based on windowing and histogram equalization of the present invention;

[0073] Figure 7 It is a CT image after a mixed Gaussian model is constructed in an optimized automatic CT image enhancement method based on Windowing and histogram equalization of the present invention;

[0074] Figure 8 It is a CT image obtained after a standardized gradient histogram is obtained in an optimized automatic CT image enhancement method based on Windowing and histogram equalization of the present invention;

[0075] Figure 9-12 It is a comparison effect diagram of the slices in tiff format obtained after the input CT image is processed by an optimized automatic CT image enhancement method based on Windowing and histogram equalization of the present invention and the slices before processing; DETAILED DESCRIPTION

[0076] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0077] An embodiment of an optimized automatic CT image enhancement method based on Windowing and histogram equalization of the present invention is as follows: Figure 1 As shown, including:

[0078] Step 1: Standardization;

[0079] The purpose of this step is to adjust the CT image pixel values ​​to a range between 0 and 1. This normalization process ensures consistency and comparability across different scans and modalities, and facilitates accurate quantitative analysis and visual assessment of image features. By uniformly scaling pixel intensity, the effects of image acquisition settings or equipment differences can be eliminated, providing a standardized basis for subsequent image processing and diagnostic interpretation;

[0080] The standardized method is:

[0081] (1) Read files: Read NIfTI files and use the nibabel library to load medical image files in NIfTI format (*.nii.gz), and then store the data as nibabel image objects;

[0082] (2) Rotate the image: Extract the image data from the nibabel image object as a numpy array, rotate the array 90 degrees to standardize the image orientation and ensure spatial orientation consistency, and finally return the rotated numpy array;

[0083] (3) Calculate the maximum and minimum values: Get I(x,y) as the original intensity value of the pixel with coordinates (x,y) in the CT image, and calculate the minimum I in the entire image. min and maximum I max Strength value;

[0084]

[0085] (4) Standardized image: Then use the minimum-maximum normalization formula to normalize the intensity I(x,y) of each pixel to I norm (x,y);

[0086]

[0087] Formula explanation:

[0088] I min Make sure the lowest intensity value in the image is mapped to 0.

[0089] I max Make sure the highest intensity value in the image is mapped to 1.

[0090] Numerator I(x,y)-I min Calculate I(x,y) relative to I min The offset of

[0091] Denominator I max -I min The intensity range is scaled to the interval [0,1].

[0092] By applying standardization, the effect visualization is obtained as follows Figure 3As shown, each pixel in the CT image is converted into a standardized intensity value, thereby improving the quality and effect of subsequent image processing tasks.

[0093] Step 2: Inhibit the bones;

[0094] The purpose of this step is to limit the pixel intensity above the specified threshold in the CT image after normalization, reducing the prominence of bone structures;

[0095] The methods of inhibiting bones are:

[0096] (1) Determine the bone strength threshold: Set a threshold T based on experience and statistical characteristics of the image. This threshold T is used to distinguish between bones (high strength) and soft tissue (low strength).

[0097] T=Threshold for bone suppression

[0098] (2) Suppress bone strength: For each pixel I(x, y) in the image, if its intensity value is higher than the threshold T, its intensity value is limited to the threshold T. The specific suppression formula is as follows:

[0099]

[0100] In the formula, I suppressed (x, y) is the intensity value after suppression, I(x, y) is the original intensity value of the pixel with coordinates (x, y) in the CT image, and T is the threshold for distinguishing between bones (high intensity) and soft tissues (low intensity).

[0101] Formula explanation:

[0102] When the value of I(x,y) does not exceed the threshold T, the pixel value remains unchanged.

[0103] When the value of I(x,y) exceeds the threshold T, the pixel value is limited to T.

[0104] By suppressing bone processing, the effect is visualized as follows Figure 4 As shown, the bone pixel intensity above the threshold T in the standardized CT image is suppressed, thereby weakening the prominence of the bone structure in the image and making other structures such as soft tissue and blood vessels more obvious.

[0105] Step 3: Use CLAHE to enhance contrast;

[0106] The purpose of this step is to enhance the contrast of CT images by applying contrast-limited adaptive histogram equalization (CLAHE) after bone suppression;

[0107] The methods to enhance contrast are:

[0108] (1) Divide the image into sub-images: Divide the original image I into several non-overlapping sub-images I i,j , where i,j represent the row and column indices of the subgraph.

[0109]

[0110] In the formula, I is the original image, I i,j To divide the original image into several non-overlapping sub-images, i, j represent the row and column indexes of the sub-image, m×n is set to the size of the original image, and a×b is set to the size of each sub-image.

[0111] (2) Calculate the histogram of each subgraph: For each subgraph I i,j Calculate its grayscale histogram H i,j (k), where k represents the gray level (0 to 255):

[0112]

[0113] In the formula, H i,j (k) is sub-graph I i,j Grayscale histogram, k represents the grayscale level (0 ~ 255), where δ is the Kroneckerdelta function, when I(x, y) = k, δ = 1, otherwise δ = 0.

[0114] (3) Limiting the histogram: To avoid excessive noise enhancement, limit the frequency H of each gray level in the histogram i,j (k) Does not exceed a set contrast limit threshold clipLimit.

[0115]

[0116] Evenly distribute the part exceeding clipLimit to other gray levels:

[0117]

[0118] Where clipLimit is the set contrast limit threshold.

[0119] (4) Calculate the cumulative distribution function (CDF): Map the original grayscale value to the enhanced grayscale value through the normalized cumulative distribution function:

[0120]

[0121] Enhanced gray value By looking up the table:

[0122]

[0123] In the formula, is the enhanced grayscale value, obtained by looking up the table.

[0124] (5) Image reorganization: All sub-images are reassembled Combine back to original image size:

[0125]

[0126] In the formula, I enhanced is the enhanced original image combined from all sub-images, For sub-graph.

[0127] By applying CLAHE, the visualization of subtle tissue differences and vascular structures was improved. Figure 5 As shown in the figure, the local contrast and detail visibility of CT images are significantly improved, blood vessels and other details are clearer, and the overall image does not over-amplify noise, maintaining visual quality. This technology ensures that important diagnostic information is not obscured by uneven lighting or image artifacts. The enhanced contrast helps to interpret anatomical features more clearly, helping radiologists identify abnormalities and make clinical decisions based on improved image clarity and details.

[0128] Step 4: Adjust window width and window position;

[0129] The purpose of this step is to change the intensity range displayed on the CT image by adjusting the window width and window position after applying CLAHE;

[0130] The window width (WW) determines the grayscale range of the displayed image. A narrower window width increases contrast, making the details of the image clearer and more prominent, and is suitable for situations where a local area needs to be carefully observed. Conversely, a wider window width reduces contrast, making the image appear smoother, and is suitable for observing a large range of structures.

[0131] The window level (WL) determines the center value of the grayscale range, that is, the adjusted brightness level. By changing the window level, pixels of different intensity ranges can be adjusted to the middle of the grayscale, thereby better displaying the area of ​​interest. For example, by raising the window level, high-intensity bone structures can be better displayed; lowering the window level helps to highlight low-intensity soft tissues.

[0132] The method of adjusting the window width and window level is to set the upper and lower limits of the window width and window level, and remap the intensity value of each pixel. Pixel values ​​below the lower limit are displayed in black, and those above the upper limit are displayed in white, while pixel values ​​between the upper and lower limits are linearly mapped to the grayscale range.

[0133] By adjusting the window width and window position, the effect can be visualized as follows Figure 6As shown in the figure, the contrast and brightness of CT images are improved, effectively highlighting specific structures such as soft tissue and blood vessels, thereby improving the diagnostic value of the image.

[0134] Step 5: Construct a mixed Gaussian model;

[0135] The purpose of this step is to construct a Gaussian mixture model (GMM) based on the gradient distribution;

[0136] GMM is a powerful probabilistic model for representing multiple Gaussian components in data distribution. In image processing, GMM can be used to model the distribution of gradients, thereby identifying different types of image features based on gradient features. By modeling gradients, GMM can help distinguish different areas in an image, such as edges, flat areas, and areas with rich textures.

[0137] The method to construct a mixed Gaussian model is:

[0138] (1) Gradient calculation: First, the gradient of the image is calculated to obtain the gradient magnitude G and direction θ of each pixel. The gradient magnitude G and direction θ are usually calculated using the Sobel operator or other differential operators.

[0139]

[0140] (2) Data preparation: The gradient information (such as gradient magnitude and direction) of each pixel in the image is formed into data points and prepared for input into the GMM. For each pixel (x, y), the corresponding gradient data point is (G(x, y), θ(x, y)).

[0141] (3) Model initialization: Select the number of Gaussian components k and initialize the parameters of GMM, including the mean, covariance matrix and mixing coefficient of each component. Initialization can be done by random selection or using the results of K-means clustering as the initial value.

[0142] (4) Expectation Maximization (EM) algorithm: Use the EM algorithm to iteratively optimize the parameters of GMM. The EM algorithm consists of two main steps:

[0143] 1E step (expected step): Calculate the probability (posterior probability) that each data point belongs to each Gaussian component:

[0144]

[0145] In the formula, π k is the mixing coefficient of the kth Gaussian component, Ν(x i μ k ,Σ k ) is the probability density function of the kth Gaussian component.

[0146] 2M steps (maximization step): Update the parameters of GMM to maximize the likelihood function:

[0147]

[0148] In the formula, μ k is the mean of the kth Gaussian component, Σ k is the covariance matrix, γ ik is the probability that the kth data point belongs to the kth component, and N is the total number of data points.

[0149] (5) Model evaluation and adjustment: The EM algorithm is iterated until convergence. After the model converges, the performance of the model can be evaluated and the number of Gaussian components k can be adjusted or the parameters can be reinitialized as needed.

[0150] By constructing a mixed Gaussian model, the effect is visualized as follows Figure 7 As shown in the figure, the gradient features of different regions in the CT image can be accurately captured, which is helpful for subsequent image segmentation and feature recognition. GMM has a wide range of applications in image processing, especially in medical image analysis, where it can effectively distinguish different types of tissues and lesion areas.

[0151] Step 6: Get the standardized gradient histogram;

[0152] The purpose of this step is to normalize the gradient information and further standardize the edge information in the image so that the edge intensity of the CT image is represented within a uniform range;

[0153] The method to obtain the standardized gradient histogram is:

[0154] (1) Gradient calculation: First, the gradient value of each pixel in the image is calculated. These gradient values ​​are obtained through a difference operator (Sobel operator) and represent the change of each pixel in the horizontal and vertical directions.

[0155] (2) Gradient strength and direction: Calculate the gradient strength and direction of each pixel. The gradient strength indicates the magnitude of the change in pixel value, while the gradient direction indicates the direction of the change. This information is crucial for understanding edges and boundaries in images.

[0156] (3) Standardization: The calculated gradient values ​​are standardized so that they fall within a uniform range (e.g., between 0 and 1). The purpose of standardization is to eliminate the differences in gradient strength between different images so that the analysis results are comparable.

[0157] (4) Histogram calculation: The normalized gradient values ​​are assigned to different gradient buckets to form a gradient histogram. The gradient histogram counts the frequency distribution of different gradient values ​​and provides a global view of the gradient features in the image.

[0158] (5) Histogram smoothing: In order to reduce the impact of noise on the gradient histogram, the histogram can be smoothed. This step helps to show the distribution trend of the gradient value more clearly.

[0159] By obtaining the normalized gradient histogram, we can get a histogram showing the normalized gradient intensity frequency distribution in the entire image. Figure 8 As shown in the figure, this processing makes the edge features of the CT image more consistent. This not only improves the effect of subsequent processing steps (such as clustering and segmentation), but also enhances the understanding of the structure of the CT image. In medical image analysis, the standardized gradient histogram can also help accurately identify and analyze complex anatomical structures and lesion areas, improving the accuracy and efficiency of diagnosis.

[0160] Step 7: Smoothing;

[0161] The purpose of this step is to further reduce the noise and enhance the clarity of the CT images using Gaussian smoothing;

[0162] Gaussian smoothing helps reduce the effects of noise in CT images, thereby improving the visual clarity of structures. Gaussian smoothing processes images through a convolution operation, using a Gaussian kernel to weight the values ​​of pixels in a neighborhood to achieve the effect of image smoothing.

[0163] The smoothing method is:

[0164] (1) Definition of Gaussian kernel: Gaussian kernel G(x,y) is a two-dimensional normal distribution function, defined as:

[0165]

[0166] Where σ is the standard deviation and (x,y) are the coordinates of the kernel.

[0167] (2) Discretization of Gaussian kernel: In actual operation, the Gaussian kernel needs to be discretized into a (2k+1)×(2k+1) matrix with the center point at (0,0) and the range of -k≤x,y≤k. A common choice is k≈3σ to ensure that the value of the Gaussian kernel is close enough to zero:

[0168]

[0169] Where i, j are the indices of the Gaussian kernel matrix, and k is the center index of the kernel (i.e. ).

[0170] (3) Image convolution: Convolve the Gaussian kernel with the original image to generate a smoothed image. The convolution operation is achieved by taking a weighted average of the neighborhood pixels around each pixel:

[0171]

[0172] In the formula, I smooth (x,y) is the smoothed image, G(i,j) is the weight of the Gaussian kernel,

[0173] I(xi,yj) is the pixel value of the original image, and k is the size of the kernel.

[0174] By applying Gaussian smoothing, the noise in CT images is effectively reduced while maintaining important structures and details in the images, making subsequent processing and analysis more reliable.

[0175] Step 8: Visualize using maximum intensity projection;

[0176] The purpose of this step is to visualize the segmented or enhanced image using the maximum intensity projection (MIP) method;

[0177] Maximum Intensity Projection (MIP) is a powerful image visualization technique widely used in medical imaging, especially in CT and MRI scans. MIP provides comprehensive visualization of structures with enhanced contrast and clarity by projecting the highest intensity values ​​in the image data along a selected viewing direction.

[0178] The method for visualization using maximum intensity projection is;

[0179] (1) Data preparation: Before performing MIP, CT image data has been preprocessed, segmented, or enhanced.

[0180] (2) Projection direction selection: MIP can be projected from different directions, such as transverse, longitudinal, coronal, or any angle. Selecting the appropriate projection direction can highlight specific structures or areas. For example, projection along the direction of blood vessels can clearly display the vascular network.

[0181] (3) Maximum intensity projection: In the selected projection direction, the MIP algorithm scans the image data layer by layer and finds the highest intensity value along each projection line. These highest intensity values ​​are projected onto a two-dimensional plane to form the final MIP image. In this way, MIP can effectively display the brightest part of the image, which usually corresponds to the anatomical structure or lesion area of ​​interest.

[0182] (4) Image generation and adjustment: After the MIP image is generated, further adjustments and optimizations can be performed. For example, the brightness and contrast can be adjusted to enhance the visual effect. In addition, color mapping can be applied to highlight different intensity values.

[0183] (5) Visualization and analysis: The final MIP image already provides comprehensive visualization with enhanced contrast and clarity. If you want to analyze each slice individually, you can also save the MIP image in TIFF format to obtain the following: Fig.10 The effect shown. In this way, complex anatomical structures and lesion areas become clearer and easier to identify. This is of great significance for diagnosis and treatment planning.

[0184] By using maximum intensity projection for visualization, not only the visualization of the image is improved, but also the accuracy and efficiency of diagnosis are enhanced. In addition, MIP images can also be used for preoperative planning and postoperative evaluation to help doctors better understand the patient's anatomical structure and lesions. Through MIP technology, complex three-dimensional medical image data becomes more intuitive and easy to interpret, thus playing an important role in medical diagnosis and treatment.

[0185] After executing the above process on the CT image, the following Figure 9-11 effect.

[0186] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An optimized automatic CT image enhancement method based on Windowing and histogram equalization, characterized in that: The method includes: Normalization: The intensity value of each pixel of the CT image is normalized to the range of [0, 1] using the minimum-maximum normalization formula; Suppress bones: suppress pixels with high intensity bone structures by applying a threshold; Contrast enhancement using CLAHE: Apply Contrast-Limited Adaptive Histogram Equalization (CLAHE); Window width and window level adjustment: adjust the window width (WW) and window level (WL) settings; Construct a mixed Gaussian model: Construct a mixed Gaussian model (GMM) based on the gradient distribution; Get the normalized gradient histogram; Smoothing: Apply Gaussian smoothing; Visualization using Maximum Intensity Projection: The enhanced or segmented images are visualized using the Maximum Intensity Projection (MIP) technique.

2. The optimized automatic CT image enhancement method based on Windowing and histogram equalization according to claim 1, characterized in that: The standardized method is: Minimum-maximum normalization formula: In the formula, I norm (x, y) is the normalized intensity value after standardization, I(x, y) is the original intensity value of the pixel with coordinates (x, y) in the CT image, and I min is the minimum intensity value in the image, I max is the maximum intensity value in the image.

3. The optimized automatic CT image enhancement method based on Windowing and histogram equalization according to claim 1, characterized in that: The method of suppressing pixels of high-intensity bone structures by applying a threshold is: Inhibition formula: T=Threshold for bone suppression In the formula, I suppressed (x, y) is the intensity value after suppression, I(x, y) is the original intensity value of the pixel with coordinates (x, y) in the CT image, and T is the threshold for distinguishing between bones (high intensity) and soft tissues (low intensity).

4. The optimized automatic CT image enhancement method based on Windowing and histogram equalization according to claim 1, characterized in that: The method of applying Contrast Limited Adaptive Histogram Equalization (CLAHE) is: Contrast Enhancement using CLAHE formula: In the formula, I is the original image, I i,j To divide the original image into several non-overlapping sub-images, i, j represent the row and column indexes of the sub-image, m×n is set to the size of the original image, and a×b is set to the size of each sub-image. In the formula, H i,j (K) is sub-graph I i,j Grayscale histogram, k represents the grayscale level (0-255), δ is the Kronecker delta function, when I(x,y)=k, δ=1, otherwise δ=0. In the formula, clipLimit is the set contrast limit threshold, and the part exceeding clipLimit is evenly distributed to other gray levels. In the formula, is the enhanced grayscale value, obtained by looking up the table. In the formula, I enhanced is the enhanced original image combined from all sub-images, For sub-graph.

5. The optimized automatic CT image enhancement method based on Windowing and histogram equalization according to claim 1, characterized in that: The method for adjusting the window width and window level is: Set the upper and lower limits of the window width and window level to remap the intensity value of each pixel. Pixel values ​​below the lower limit are displayed in black, those above the upper limit are displayed in white, and pixel values ​​between the upper and lower limits are linearly mapped to the grayscale range.

6. The optimized automatic CT image enhancement method based on Windowing and histogram equalization according to claim 1, characterized in that: The method for constructing the mixed Gaussian model is: Where G is the gradient size of the pixel and θ is the direction of each pixel. Prepare the gradient data of each pixel (G(x,y),θ(x,y)) to be input into GMM (Gaussian Mixture Model); initialize GMM and use the expectation maximization (EM) algorithm to iteratively optimize the parameters of GMM; Expectation-Maximization (EM) algorithm: E step (expected step): In the formula, π k is the mixing coefficient of the kth Gaussian component, Ν(x i |μ k ,Σ k ) is the probability density function of the kth Gaussian component. M step (maximization step): In the formula, μ k is the mean of the kth Gaussian component, ∑ k is the covariance matrix, γ ik is the probability that the kth data point belongs to the kth component, and N is the total number of data points.

7. The optimized automatic CT image enhancement method based on Windowing and histogram equalization according to claim 1, characterized in that: The method for obtaining the standardized gradient histogram is: The Sobel operator is used to calculate the gradient value of each pixel in the image, and then the gradient strength and direction of each pixel are calculated. The calculated gradient values ​​are then standardized and assigned to different gradient intervals to form a gradient histogram, which is then smoothed.

8. The optimized automatic CT image enhancement method based on Windowing and histogram equalization according to claim 1, characterized in that: The smoothing method: Gaussian smoothing formula: Where, Gaussian kernel G(x,y) is a two-dimensional normal distribution function, σ is the standard deviation, and (x,y) is the coordinate of the kernel. Where i, j are the indices of the Gaussian kernel matrix, and k is the center index of the kernel (i.e. In the formula, I smooth (x,y) is the smoothed image, G(i,j) is the weight of the Gaussian kernel, I(xi,yj) is the pixel value of the original image, and k is the size of the kernel.

9. The optimized automatic CT image enhancement method based on Windowing and histogram equalization according to claim 1, characterized in that: The method described for visualization using maximum intensity projection: Select the projection direction, use MIP (Maximum Intensity Projection) for the data source, obtain the processed image, and adjust the brightness and contrast or apply color mapping when displaying the image.

Citation Information

Cited By

  • Intelligent orthopedic CT image processing system

    CN120912611A

  • X-ray image enhancement method and system and storage medium

    CN121788392A

  • An X-ray image enhancement method, system, and storage medium

    CN121788392B

  • CT image enhancement method and device based on global consistency optimization

    CN122089577A