Image CT (Computed Tomography) image enhancement method and system

Abstract

The invention discloses an image CT (Computed Tomography) image enhancement method and system. The image enhancement method comprises the following steps: acquiring a CT image; preprocessing the CT image to obtain a feature vector; clustering different tissue structures in the CT image based on the feature vectors to obtain tissue structure images corresponding to different categories; and performing saturation transparent adjustment on the tissue structure image to obtain an adjusted CT image. The visual difference of the tissue structure can be enhanced from multiple angles, the overall quality and readability of the CT image are improved, and clear and accurate information support is provided for image observation and diagnosis of doctors.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and in particular relates to a method and system for enhancing a CT image. Background Art

[0002] CT image data contains a wealth of information about human tissue structures, such as bones and soft tissues. However, the grayscale values ​​of different tissue structures on CT images are slightly different, and the contrast is not high, resulting in blurred boundaries of tissue structures and difficulty in accurate distinction. At the same time, the overall grayscale distribution of CT images is uneven, the local contrast is low, and important tissue structures are not clearly displayed. This makes it impossible for doctors to quickly and accurately identify and locate lesion areas when observing CT images, affecting diagnostic efficiency and accuracy. Therefore, how to effectively improve the contrast and boundary clarity of different tissue structures in CT images, so that structures such as bones and soft tissues are more prominent and easier to distinguish, is a key technical problem that needs to be solved urgently. Summary of the invention

[0003] To solve the above technical problems, the present invention proposes a CT image enhancement method and system, which can enhance the visual differences of tissue structures from multiple angles, improve the overall quality and readability of CT images, and provide clear and accurate information support for doctors' image observation and diagnosis.

[0004] The present invention provides a method for enhancing CT images, comprising:

[0005] Acquire CT images;

[0006] Preprocessing the CT image to obtain a feature vector;

[0007] Based on the feature vector, clustering different tissue structures in the CT image to obtain tissue structure images corresponding to different categories;

[0008] The saturation and transparency of the tissue structure image are adjusted to obtain an adjusted CT image.

[0009] Optionally, preprocessing the CT image to obtain a feature vector includes:

[0010] Segmenting the CT image to obtain segmented tissue structure data;

[0011] According to the segmented tissue structure data, the gray value distribution characteristics of the tissue structure data are calculated to obtain a feature vector.

[0012] Optionally, preprocessing the CT image, segmenting the CT image, and obtaining the segmented tissue structure data includes:

[0013] Calculating the grayscale histogram of the CT image, and adaptively determining the optimal segmentation threshold by judging the peaks and valleys of the grayscale histogram;

[0014] The CT image data is binarized and segmented according to the optimal segmentation threshold to obtain segmented tissue structure data.

[0015] Optionally, clustering different tissue structures in the CT image based on the feature vector to obtain tissue structure images corresponding to different categories includes:

[0016] Performing data cleaning and normalization processing on the feature vector to obtain a processed feature vector;

[0017] Based on the processed feature vectors, the K-means algorithm is used to perform clustering calculations to obtain preliminary clustering results;

[0018] The preliminary clustering results are processed by removing outliers and merging small classes to obtain clustering results;

[0019] The clustering results are mapped to tissue structure categories to obtain tissue structure images corresponding to the different categories.

[0020] Optionally, removing outliers and merging small classes from the preliminary clustering results, and obtaining the clustering results includes:

[0021] Using the DBSCAN algorithm to mark the preliminary clustering results, and remove the marked noise points;

[0022] The remaining clustering results are clustered again using an agglomerative hierarchical clustering algorithm to generate a cluster dendrogram, and the cluster dendrogram is cut according to a preset distance threshold to determine a number of clusters;

[0023] Calculate the number of data points contained in each cluster and obtain a set of small-scale clusters;

[0024] Based on the set of small-scale clusters, calculate the distance between each small-scale cluster and the center point of other clusters, merge the small-scale clusters, and obtain the clustering result after merging the small-scale clusters;

[0025] Based on the clustering results after merging small-scale clusters, the silhouette coefficient of each data point is calculated to obtain the best clustering result.

[0026] Optionally, performing saturation transparency adjustment on the tissue structure image to obtain the adjusted CT image includes:

[0027] Determining adjustment parameters corresponding to the tissue structure image according to a preset tissue structure type;

[0028] Convert the tissue structure image to the HSV color space, and use the corresponding adjustment parameters for the three channels H, S, and V to adjust the values ​​of the saturation S channel and the lightness V channel;

[0029] The adjusted HSV image is converted back to the RGB color space to obtain an adjusted tissue structure RGB image.

[0030] Optionally, the method further includes: performing three-dimensional visual display on the adjusted CT image.

[0031] Optionally, performing three-dimensional visualization display on the adjusted CT image includes:

[0032] Using a three-dimensional reconstruction algorithm, the CT image is converted into three-dimensional model data to obtain a three-dimensional CT image model;

[0033] Surface rendering parameters are determined according to the characteristics of the three-dimensional CT image model, and the three-dimensional CT image is generated using the surface rendering parameters.

[0034] The present invention also provides an image CT image enhancement system, comprising: an image acquisition module, an image processing module, a clustering module, an image adjustment module and a visualization display module;

[0035] The image acquisition module is used to acquire CT images;

[0036] The image processing module is used to pre-process the CT image to obtain a feature vector;

[0037] The clustering module is used to cluster different tissue structures in the CT image to obtain tissue structure images corresponding to different categories;

[0038] The image adjustment module is used to adjust the saturation transparency of the tissue structure image to obtain an adjusted CT image;

[0039] The visualization display module is used to generate a three-dimensional visualization result of the CT image.

[0040] Compared with the prior art, the present invention has the following advantages and technical effects:

[0041] The present invention first pre-processes the acquired CT image data, and then adaptively selects threshold segmentation or region growing method to segment different tissue structures such as bone tissue and soft tissue in the CT image. Based on the segmented tissue structure data, the gray value distribution characteristics of each tissue structure are calculated to obtain a feature vector reflecting the gray distribution of different tissue structures. Then, a clustering algorithm is used to cluster the feature vectors to obtain category data of different tissue structures. Different saturation and transparency adjustment parameters are set for different tissue structure categories, and the visual contrast of different tissue structures is improved by pixel-level adjustment. Finally, the adjusted CT image is enhanced, and three-dimensional reconstruction, surface rendering and other technologies are used to generate intuitive and realistic CT image visualization results, which are convenient for doctors to observe and diagnose. The method proposed by the present invention effectively improves the quality and visualization effect of CT images through the organic combination of tissue structure segmentation, saturation transparency adjustment and image enhancement processing, and provides clearer and more accurate image data support for medical diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0043] Figure 1 It is a flow chart of a CT image enhancement method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0045] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0046] CT enhanced scanning is a common diagnostic technique in medical imaging. It improves the contrast of CT images and the ability to visualize specific tissues and blood vessels by injecting patients with a special substance called contrast agent. Contrast agent is usually a substance containing iodine, which has the property of high absorption of X-rays in the body, so that it can enhance the visibility of blood vessels, organs and other tissues in CT scans. By injecting contrast agent into the patient intravenously before the scan, it will be transported to various parts of the body through the blood circulation, thus producing a more obvious contrast effect compared with surrounding tissues during the scan. This contrast effect enables doctors to more clearly identify and evaluate the presence and characteristics of diseases such as lesions, tumors, infections, and vascular diseases. Enhanced scanning can also be used to evaluate the function of organs. By observing the distribution and excretion process of contrast agent in organs, doctors can evaluate the health status of organs such as kidney function and liver function, thereby guiding clinical decisions and treatment plans. In addition, in interventional treatment, enhanced scanning can help doctors accurately locate treatment targets and ensure that therapeutic substances are correctly delivered to specific areas.

[0047] Enhanced CT is a method of examination that is performed after intravenous injection of water-soluble iodine contrast agent into the suspicious area on the basis of CT plain scan. Its purpose is to increase the density of the lesion tissue and the adjacent normal tissue, improve the display rate of the lesion tissue, and help doctors observe the condition of the lesion tissue. Reasons for doing enhanced CT: 1. Enhanced CT examination can provide clearer imaging information. In the diagnosis of some diseases, ordinary CT images may not be able to clearly display or detect the lesion site, while enhanced CT images can make the lesion site more vivid and visible through the spread of contrast agents in the blood. 2. Enhanced CT examination can improve the accuracy of diagnosis. For some difficult cases or situations that require further clarification of the diagnosis, enhanced CT can help doctors more accurately determine the nature, scope and severity of the lesion in order to formulate a more reasonable treatment plan. 3. Enhanced CT examination can evaluate vascular perfusion. For some lesions, such as tumors, vascular embolism, etc., the vascular supply status related to them is very important information. Enhanced CT examination can provide detailed information about vascular perfusion and provide guidance for doctors to make accurate diagnosis and treatment. 4. Determine tumor stage: For tumor patients, enhanced CT can more accurately determine the tumor stage and help formulate appropriate treatment plans.

[0048] The basic principle of enhanced CT: Enhanced CT increases the density difference between lesion tissue and normal tissue by intravenous injection of iodine contrast agent, thereby improving the resolution and contrast of the image. After the contrast agent enters the blood vessels, the vascular development will be enhanced, which helps to distinguish vascular and non-vascular lesions.

[0049] Application scenarios and advantages of enhanced CT: 1. Improve the detection rate of lesions, especially small lesions. Through enhanced scanning, the degree of enhancement of lesions and parenchyma or surrounding tissues is often inconsistent, and the density difference between the two is large, which can be used to find lesions that were missed during plain scanning. During plain scanning, these lesions are of equal density or have very small density differences, or cannot be distinguished from vascular cross-sectional shadows. 2. Improve the ability to characterize lesions. CT plain scanning not only has a low detection rate, but also has limited ability to characterize and distinguish lesions. Dynamic enhanced scanning can improve the ability to characterize lesions according to the presence, degree, and mode or type of enhancement of lesions. It is not difficult to make a qualitative diagnosis for typical cases. 3. Improve the accuracy of tumor staging. In patients who have been diagnosed with malignant tumors, the purpose of enhanced scanning is to improve the accuracy of tumor staging or to determine the possibility of surgical resection of the tumor. Enhanced scanning is also very important for the identification of vascular and non-vascular lesions, such as the identification of blood vessels and small lymph nodes. It is often difficult to distinguish lymph nodes and blood vessels in the hilum, neck, pelvis and other parts of the image. Through enhanced scanning, the density difference between high-density blood vessels and lymph nodes with unclear enhancement is large, which is easy to distinguish. 4. Improve image resolution and contrast: Enhanced CT can make blood vessels and lesions more obvious, which helps to judge the condition more accurately. Provide more information: In addition to showing the size, location and morphology of the lesion, it can also observe the blood supply of the lesion, the degree of involvement of the surrounding tissues, etc., providing more comprehensive information for clinical diagnosis and treatment. Reduce radiation dose: Compared with traditional CT scans, enhanced CT uses a lower dose of X-rays and causes less radiation damage to patients. Improve the detection rate of lesions: Enhanced CT can scan tiny lesions that are not easy to find, making the boundary between lesion tissue and normal tissue clear. Facilitate the positioning and diagnosis of tumors: Enhanced CT can provide staging judgments for malignant tumors, accurately judge the location and degree of invasion of tumors, and help determine the scope and method of surgery.

[0050] Precautions for enhanced CT: 1. The contrast agent circulates through the blood throughout the body and is finally excreted through the kidneys. 99% of the contrast agent will be excreted from the kidneys, so patients with renal failure cannot undergo this examination. For patients with normal renal function, they should drink plenty of water within 24 hours after CT enhancement, fully hydrate, and urinate to allow the contrast agent to be excreted as soon as possible. 2. People who are allergic to iodine cannot undergo enhanced CT examination. Before doing enhanced CT, the doctor will ask about the allergy history in detail. Severe allergies are relatively rare, but once they occur, they can cause anaphylactic shock and even respiratory and cardiac arrest. Therefore, after the enhanced CT, you must observe outside the examination room for half an hour before leaving. 3. Patients who undergo abdominal and pelvic enhanced CT should empty their intestines as much as possible, fast for one night, and reduce the interference of feces and intestinal gas on imaging. Upper gastrointestinal tract contrast or barium enema should not be performed one week before CT to prevent the contrast agent and barium sulfate from remaining in the intestinal cavity and affecting the imaging. 4. If you have severe hyperthyroidism and the condition is not well controlled, try not to do enhanced CT, because the injected iodine-containing contrast agent may aggravate the hyperthyroidism. Since contrast agents are injected into the body by a high-pressure syringe, patients with severe pulmonary insufficiency may suffer from heart failure and respiratory failure if contrast agents are injected quickly. Therefore, a full evaluation is required before performing enhanced CT.

[0051] By comprehensively using a variety of technical means such as image segmentation, feature extraction, clustering, saturation and transparency adjustment, and image enhancement, the visual differences of tissue structures are enhanced from multiple angles, the overall quality and readability of CT images are improved, and clear and accurate information support is provided for doctors' image observation and diagnosis. At the same time, quantitative evaluation indicators are established to objectively evaluate the effect of image quality improvement to guide the optimization and improvement of the algorithm.

[0052] This embodiment proposes a method for enhancing CT images. Figure 1 As shown, the specific steps include:

[0053] Acquire CT images;

[0054] Preprocess the CT image to obtain the feature vector;

[0055] Based on the feature vector, different tissue structures in the CT image are clustered to obtain tissue structure images corresponding to different categories;

[0056] The saturation and transparency of the tissue structure image are adjusted to obtain the adjusted CT image.

[0057] Specifically, this embodiment first pre-processes the acquired CT image data, and then adaptively selects threshold segmentation or region growing method to segment different tissue structures such as bone tissue and soft tissue in the CT image. Based on the segmented tissue structure data, the gray value distribution characteristics of each tissue structure are calculated to obtain a feature vector reflecting the gray distribution of different tissue structures. Then, a clustering algorithm is used to cluster the feature vectors to obtain category data of different tissue structures. Different saturation and transparency adjustment parameters are set for different tissue structure categories, and the visual contrast of different tissue structures is improved by pixel-level adjustment. Finally, the adjusted CT image is enhanced, and three-dimensional reconstruction, surface rendering and other technologies are used to generate intuitive and realistic CT image visualization results, which are convenient for doctors to observe and diagnose. The method proposed in the present invention effectively improves the quality and visualization effect of CT images through the organic combination of tissue structure segmentation, saturation transparency adjustment and image enhancement processing, and provides clearer and more accurate image data support for medical diagnosis.

[0058] Furthermore, the CT image is preprocessed to obtain the feature vector including:

[0059] Segment the CT image to obtain segmented tissue structure data;

[0060] According to the segmented tissue structure data, the gray value distribution characteristics of the tissue structure data are calculated to obtain a feature vector.

[0061] Further, the CT image is preprocessed, the CT image is segmented, and the segmented tissue structure data is obtained, including:

[0062] Calculate the grayscale histogram of the CT image, and adaptively determine the optimal segmentation threshold by judging the peaks and valleys of the grayscale histogram;

[0063] The CT image data is binarized and segmented according to the optimal segmentation threshold to obtain the segmented tissue structure data.

[0064] Specifically, CT image data is acquired, and preprocessing operations such as image enhancement and denoising are performed on the CT image data to obtain preprocessed CT image data; based on the preprocessed CT image data, threshold segmentation is adaptively selected, wherein the threshold segmentation method selects the best threshold based on the image grayscale histogram, and binarizes the CT image data, and the region growing method starts from the seed point and classifies adjacent pixels into the same region according to the grayscale similarity criterion; if the threshold segmentation method is selected, the grayscale histogram is calculated for the preprocessed CT image data, and the best segmentation threshold is adaptively determined by judging the peaks and troughs of the grayscale histogram, and the CT image data is binarized and segmented according to the threshold to obtain segmented bone tissue, soft tissue and other different tissue structure data.

[0065] Furthermore, based on the feature vector, different tissue structures in the CT image are clustered to obtain tissue structure images corresponding to different categories, including:

[0066] Perform data cleaning and normalization on the feature vector to obtain the processed feature vector;

[0067] Based on the processed feature vectors, the K-means algorithm is used to perform clustering calculations to obtain preliminary clustering results;

[0068] Remove outliers and merge small classes from the preliminary clustering results to obtain clustering results;

[0069] The clustering results are mapped to the tissue structure categories to obtain tissue structure images corresponding to different categories.

[0070] Specifically, for each tissue structure area, the grayscale mean and grayscale variance of all pixels in the area are calculated to construct the grayscale distribution feature vector of the area, which includes the grayscale mean eigenvalue and the grayscale variance eigenvalue; cluster analysis is performed on the grayscale distribution feature vectors of all tissue structure areas, and the K-means clustering algorithm is used to divide the tissue structure areas into several categories according to the similarity of grayscale distribution characteristics; the grayscale mean eigenvalue and grayscale variance eigenvalue of each cluster center are calculated to obtain the grayscale distribution feature template vector of each type of tissue structure; if the new If the distance between the grayscale distribution feature vector of the tissue structure area and a certain type of template vector is the shortest, then the area is judged to belong to this type of tissue structure; for different types of tissue structures with obvious differences in grayscale distribution, Gaussian mixture models of their grayscale distribution are constructed respectively, and the essential differences in the grayscale distribution of different tissue structures are characterized by the parameter differences of the Gaussian mixture model; for new tissue structure areas to be identified, their grayscale distribution feature vectors are extracted, and the distance is measured with the feature template vectors of different types of tissue structures respectively, and combined with the Gaussian mixture model of its grayscale distribution, the tissue structure category to which it belongs is comprehensively judged.

[0071] Furthermore, the preliminary clustering results are processed by removing outliers and merging small classes, and the clustering results obtained include:

[0072] The DBSCAN algorithm is used to mark the preliminary clustering results and remove the marked noise points;

[0073] The remaining clustering results are clustered again using an agglomerative hierarchical clustering algorithm to generate a cluster dendrogram, and the cluster dendrogram is cut according to a preset distance threshold to determine a number of clusters;

[0074] Calculate the number of data points contained in each cluster and obtain a set of small-scale clusters;

[0075] Based on the set of small-scale clusters, the distance between each small-scale cluster and the center point of other clusters is calculated, and the small-scale clusters are merged to obtain the clustering results after merging the small-scale clusters;

[0076] Based on the clustering results after merging small-scale clusters, the silhouette coefficient of each data point is calculated to obtain the best clustering result.

[0077] Specifically, step 1: use the DBSCAN algorithm to perform preliminary clustering on the data set, obtain the preliminary clustering results and the cluster ID to which each data point belongs. Set the preset neighborhood radius parameter Eps and the minimum sample number parameter MinPts, and determine the core objects, boundary points, and noise points based on these two parameters.

[0078] Step 2: For the preliminary clustering results obtained in step 1, if a data point is marked as a noise point by the DBSCAN algorithm, then remove the data point from the data set to obtain a data set after noise removal.

[0079] Step 3: Get the data set after removing noise in step 2, cluster the remaining data points again using the agglomerative hierarchical clustering algorithm, generate a cluster dendrogram, and cut the dendrogram according to the preset distance threshold to determine each cluster.

[0080] Step 4: Traverse each cluster determined in step 3 and calculate the number of data points contained in each cluster. If the number of data points in a cluster is less than the preset minimum cluster size threshold, mark the cluster as a small-scale cluster and obtain the set of all small-scale clusters.

[0081] Step 5: For each small-scale cluster in the set of small-scale clusters obtained in step 4, calculate the distance between the center points of the small-scale cluster and other clusters. If the distance between the center points of a small-scale cluster and another cluster is less than the preset merge distance threshold, then merge the small-scale cluster into the cluster with the nearest distance to obtain the clustering result after merging the small-scale clusters.

[0082] Step 6: Use the silhouette coefficient to evaluate the quality of the clustering results after merging the small-scale clusters in step 5. Calculate the silhouette coefficient of each data point and average the silhouette coefficients of all data points to get the average silhouette coefficient. If the average silhouette coefficient is higher than the preset silhouette coefficient threshold, then the clustering result is judged to be of good quality.

[0083] Step 7: Based on the evaluation results of step 6, if the quality of the clustering results is good, then output the final clustering results; otherwise, adjust the parameters Eps and MinPts of the DBSCAN algorithm, the distance threshold of the agglomerative hierarchical clustering algorithm, and the merge distance threshold, and repeat steps 1 to 6 until a high-quality clustering result is obtained.

[0084] Further, the saturation transparency of the tissue structure image is adjusted, and the adjusted CT image is obtained, including:

[0085] Determine adjustment parameters corresponding to the tissue structure image according to the preset tissue structure type;

[0086] The tissue structure image is converted to the HSV color space, and the corresponding adjustment parameters are used for the three channels H, S, and V to adjust the values ​​of the saturation S channel and the lightness V channel;

[0087] The adjusted HSV image is converted back to the RGB color space to obtain an adjusted tissue structure RGB image.

[0088] Specifically, a tissue structure image is obtained, and adjustment parameters corresponding to each tissue structure are determined according to a preset tissue structure type; the tissue structure image is converted to an HSV color space, and corresponding adjustment parameters are used for the three channels H, S, and V to adjust the values ​​of the saturation S channel and the lightness V channel; the adjusted HSV image is converted back to an RGB color space to obtain an adjusted tissue structure RGB image; according to the tissue structure type, a preset color mapping table is used to map the color of the adjusted RGB image to a preset color gamut; the mapped RGB image is quantized, and the RGB value is converted into an index color code to obtain an adjusted tissue structure index image; according to the hierarchical relationship of the tissue structure, an improved Flood Fill algorithm is used to fill the connected areas in the tissue structure index image, and the filled image is used as the final output; when using the improved Flood Fill algorithm, different seed points are set according to the hierarchy of the tissue structure, and connected areas of different hierarchies are filled to ensure the nested relationship between the upper and lower hierarchies.

[0089] Furthermore, the method also includes: performing three-dimensional visualization display on the adjusted CT image.

[0090] Further, performing three-dimensional visualization display on the adjusted CT image includes:

[0091] Using a three-dimensional reconstruction algorithm, the CT image is converted into three-dimensional model data to obtain a three-dimensional CT image model;

[0092] According to the characteristics of the three-dimensional CT image model, surface rendering parameters are determined, and the three-dimensional CT image is generated using the surface rendering parameters.

[0093] Specifically, according to the characteristics of the CT image, the adjustment range of saturation and transparency and the parameters of the enhancement processing are determined. The CT image data is converted into 3D model data by using a 3D reconstruction algorithm to obtain a 3D CT image model. According to the characteristics of the 3D CT image model, the parameters such as illumination, material, color, etc. of the surface rendering are determined, and realistic CT image visualization results are generated by surface rendering technology. According to the needs of the doctor's observation and diagnosis, the viewing angle, scale, transparency and other parameters of the CT image visualization results are adjusted to make them more intuitive and clear. If the quality of the CT image is poor, the CT image is first pre-processed by denoising and enhancing to improve the image quality, and then 3D reconstruction and surface rendering are performed. A large amount of CT image data is analyzed by a machine learning algorithm to obtain the optimal saturation, transparency adjustment range and enhancement processing parameters, as well as the optimal 3D reconstruction and surface rendering parameters. The generated CT image visualization results are compared with the doctor's diagnosis results, and the image processing and rendering technology are continuously optimized to improve the accuracy and practicality of the visualization results.

[0094] This embodiment also provides an image CT image enhancement system, including: an image acquisition module, an image processing module, a clustering module, an image adjustment module and a visualization display module;

[0095] An image acquisition module, used for acquiring CT images;

[0096] An image processing module is used to pre-process the CT image and obtain a feature vector;

[0097] A clustering module is used to cluster different tissue structures in CT images and obtain tissue structure images corresponding to different categories;

[0098] An image adjustment module is used to adjust the saturation and transparency of the tissue structure image to obtain an adjusted CT image;

[0099] The visualization display module is used to generate three-dimensional visualization results of CT images.

[0100] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for enhancing CT images, characterized in that: include: Acquire CT images; Preprocessing the CT image to obtain a feature vector; Based on the feature vector, clustering different tissue structures in the CT image to obtain tissue structure images corresponding to different categories; The saturation and transparency of the tissue structure image are adjusted to obtain an adjusted CT image.

2. The method for enhancing CT images according to claim 1, characterized in that: Preprocessing the CT image to obtain a feature vector includes: Segmenting the CT image to obtain segmented tissue structure data; According to the segmented tissue structure data, the gray value distribution characteristics of the tissue structure data are calculated to obtain a feature vector.

3. The method for enhancing CT images according to claim 2, characterized in that: Preprocessing the CT image, segmenting the CT image, and obtaining the segmented tissue structure data includes: Calculating the grayscale histogram of the CT image, and adaptively determining the optimal segmentation threshold by judging the peaks and valleys of the grayscale histogram; The CT image data is binarized and segmented according to the optimal segmentation threshold to obtain segmented tissue structure data.

4. The method for enhancing CT images according to claim 1, characterized in that: Clustering different tissue structures in the CT image based on the feature vector to obtain tissue structure images corresponding to different categories includes: Performing data cleaning and normalization processing on the feature vector to obtain a processed feature vector; Based on the processed feature vectors, the K-means algorithm is used to perform clustering calculations to obtain preliminary clustering results; The preliminary clustering results are processed by removing outliers and merging small classes to obtain clustering results; The clustering results are mapped to tissue structure categories to obtain tissue structure images corresponding to the different categories.

5. The method for enhancing CT images according to claim 4, characterized in that: The preliminary clustering results are processed by removing outliers and merging small classes, and obtaining clustering results includes: Using the DBSCAN algorithm to mark the preliminary clustering results, and remove the marked noise points; The remaining clustering results are clustered again using an agglomerative hierarchical clustering algorithm to generate a cluster dendrogram, and the cluster dendrogram is cut according to a preset distance threshold to determine a number of clusters; Calculate the number of data points contained in each cluster and obtain a set of small-scale clusters; Based on the set of small-scale clusters, calculate the distance between each small-scale cluster and the center point of other clusters, merge the small-scale clusters, and obtain the clustering result after merging the small-scale clusters; Based on the clustering results after merging small-scale clusters, the silhouette coefficient of each data point is calculated to obtain the best clustering result.

6. The method for enhancing CT images according to claim 1, characterized in that: Performing saturation transparency adjustment on the tissue structure image to obtain the adjusted CT image includes: Determining adjustment parameters corresponding to the tissue structure image according to a preset tissue structure type; Convert the tissue structure image to the HSV color space, and use the corresponding adjustment parameters for the three channels H, S, and V to adjust the values ​​of the saturation S channel and the lightness V channel; The adjusted HSV image is converted back to the RGB color space to obtain an adjusted tissue structure RGB image.

7. The method for enhancing CT images according to claim 1, characterized in that: The method further includes: performing three-dimensional visual display on the adjusted CT image.

8. The method for enhancing CT images according to claim 7, characterized in that: The three-dimensional visualization display of the adjusted CT image includes: Using a three-dimensional reconstruction algorithm, the CT image is converted into three-dimensional model data to obtain a three-dimensional CT image model; Surface rendering parameters are determined according to the characteristics of the three-dimensional CT image model, and the three-dimensional CT image is generated using the surface rendering parameters.

9. A CT image enhancement system, characterized in that: include: Image acquisition module, image processing module, clustering module, image adjustment module and visualization display module; The image acquisition module is used to acquire CT images; The image processing module is used to pre-process the CT image to obtain a feature vector; The clustering module is used to cluster different tissue structures in the CT image to obtain tissue structure images corresponding to different categories; The image adjustment module is used to adjust the saturation transparency of the tissue structure image to obtain an adjusted CT image; The visualization display module is used to generate a three-dimensional visualization result of the CT image.

Images