A method of CTA black-blood imaging based on material separation

Abstract

The application discloses a kind of CTA black blood imaging method based on substance separation.It includes the following steps, step one: obtaining several images of human blood vessel CT enhancement;Obtain n space registered reference medical images, constitute n-dimensional gray space of substance distribution;Step two: in step one, the blood information on image is extracted by delineating the region of interest ROI in the vessel of image;Step three: for two different levels of energy image and ROI, find out the component without statistical difference with ROI area by using gray space distribution and suppress;Step four: the black blood image corresponding to the target suppression is obtained, and the vessel wall condition is highlighted.The application has the advantages of inhibiting blood signal, eliminating blood vessel pulsation artifact, reducing examination cost and shortening examination time.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a CTA black blood imaging method based on substance separation. Background Art

[0002] Related research has shown a significant link between arterial stiffness and cardiovascular diseases such as hypertension and atherosclerosis and stroke. This is because as the elasticity of the arterial wall decreases, the arterial compliance of the blood vessels decreases, leading to increased blood pressure and an increased risk of cardiovascular disease. Therefore, evaluating arterial wall properties is of great clinical significance for the prevention, monitoring, and treatment of related cardiovascular diseases.

[0003] In clinical practice, computed tomography angiography (CTA) is primarily used to study the characteristics of human vascular walls. However, traditional CTA imaging cannot distinguish between the inner and outer walls of blood vessels, and therefore cannot provide accurate information about the vessel walls. This can easily lead to missed or misdiagnosed minor lesions. Therefore, black blood imaging technology offers significant advantages in highlighting vessel wall information. CTA black blood imaging technology suppresses blood flow intensity during CT angiography, resulting in a dark, low-density blood image, thereby highlighting the vessel walls.

[0004] Currently, black blood imaging can only be achieved through magnetic resonance imaging (MRI). Its main principle is to use the "flow void effect" to transform high-speed blood flow into a "black signal," highlighting the vessel wall condition. This offers significant advantages in imaging arterial lesions and plaques. However, in MRI, many large blood vessels pulsate, resulting in vascular pulsation artifacts. Therefore, if the blood signal could be suppressed, these artifacts could be eliminated. However, blood flow in the human body is highly complex, and simply using the flow void effect cannot always guarantee 100% suppression of the blood flow void signal. Furthermore, black blood imaging is not possible in small and medium-sized vessels with slow flow. Furthermore, due to the lengthy MRI scan times, CTA scans are preferred in emergency situations to assess vascular condition. Therefore, black blood imaging cannot be performed during emergency examinations, making it easy for physicians to miss or misdiagnose subtle lesions in the vessel wall.

[0005] Therefore, it is necessary to develop a CT-based black blood technology to suppress blood signals, eliminate vascular pulsation artifacts, reduce examination costs, be widely used in blood vessels throughout the body, and shorten examination time for "black blood" imaging. Summary of the Invention

[0006] The purpose of the present invention is to provide a method for CTA black blood imaging based on material separation, which suppresses blood signals, eliminates vascular pulsation artifacts, reduces examination costs, shortens examination time (compared to MR examination methods, the present invention significantly shortens CT examination time and can be used for emergency vascular examinations), and increases applicability (the present invention is not only suitable for evaluating lesions in the heart wall and arterial wall, but also for evaluating small and medium-sized blood vessels with slow flow rates and for evaluating blood vessels throughout the body).

[0007] In order to achieve the above object, the technical solution of the present invention is: a CTA black blood imaging method based on substance separation, which is different in that it includes the following steps:

[0008] Step 1: Acquire several enhanced CT images of human blood vessels;

[0009] Obtain n (n ≥ 1) spatially registered reference medical images (referring to spectral CT angiography images) to construct an n-dimensional grayscale space of material distribution. This step is used to obtain the scanned original CTA image.

[0010] Step 2: Delineate a region of interest (ROI) within the blood vessels in the image from step 1, and extract blood information from the image from step 1. The extracted blood information primarily includes the range of blood density values ​​(CT values, HU). This step is used to delineate the ROI within the target blood vessel and obtain a .nii file of the ROI. In the next step, this file is imported into Black Blood Software (this software is part of the present invention) along with the original CTA image in .nii format, thereby outputting a black blood image.

[0011] Step 3: For the two images and ROIs at different energy levels, the grayscale spatial distribution is used to identify components that are not statistically different from the ROI region and suppress them. Different organs and tissues in the human body appear with different densities on CT. The present invention utilizes the density difference between blood and blood vessel walls to suppress high-density blood and highlight the lower-density blood vessel walls, thereby facilitating a more detailed assessment of blood vessel wall lesions by physicians. The present invention utilizes two images at different energy levels to obtain complete spectral CT information.

[0012] Step 4: Obtain a black blood image suppressed by the corresponding target to highlight the condition of the blood vessel wall. The black blood image in the present invention can better highlight the blood vessel wall, thereby facilitating a more detailed assessment of blood vessel wall lesions by physicians, which is of great significance for clinical diagnosis.

[0013] In the above technical solution, in step 1, the spatially registered images include images of different energy levels (images of any different energy levels are sufficient, and images of energy spectrum CT scans can generate images of many energy levels (40-100kev), and two common ones can be selected, such as 55kev and 100kev).

[0014] In the above technical solution, in step 1, the specific method for obtaining several images of human blood vessels enhanced by CT is:

[0015] S1.1: Use a spectral or dual-energy CT machine to acquire enhanced medical images (the enhanced medical images may be one or more).

[0016] S1.2: Register all reference images (i.e., the n (n>=1) spatially registered reference medical images obtained in step 1) so that the tissue distributions of the two (i.e., images of different energy levels) correspond pixel by pixel.

[0017] In the above technical solution, in step 2, a region of interest (ROI) is outlined within the target blood vessel, an ROI file that is spatially consistent with the reference medical image (i.e., the reference medical image in step 1) is obtained, and the substance to be separated is determined.

[0018] In the above technical solution, in step 3, the Gaussian mixture model (GMM) method is used to distinguish the n-dimensional grayscale space of the material distribution. At the same time, the grayscale information of the material in the ROI file area is referred to determine its category. Finally, the probability of each point in the grayscale space belonging to the category is calculated.

[0019] The specific method to obtain the probability that each point in the gray space belongs to this class is:

[0020] S31: Empirically estimate the number of components in the medical image as a parameter of the GMM model;

[0021] S32: Based on the existing ROI, extract the grayscale distribution in this area, and then bring it into the GMM model to obtain the category of the material in this ROI area; use the GMM model to calculate the probability that all grayscales belong to this category.

[0022] In the above technical solution, in step three, by setting an empirical threshold alpha, information in the real space of points in the grayscale space whose probability of belonging to the same type of ROI is greater than alpha is retained.

[0023] In the above technical solution, in step 4, a medical image displaying the blood components to be suppressed is obtained, and the spatial information obtained in step 3 is brought into the image. The grayscale of the voxel points (voxel point refers to the smallest unit of spatial distribution of medical image data) in these spaces is processed (such as scaling and translation, or using some suppression functions such as square root function and exponential function) so that the grayscale of voxels similar to the components of the ROI in step 2 is darkened and suppressed, thereby highlighting statistically significant tissue boundary information and obtaining a suppressed CTA black blood image;

[0024] The specific method is:

[0025] S4.1: After obtaining the spatial information from step 3, the spatial coordinates of the voxel point of this type are obtained, i.e., the index subscript of the grayscale matrix of the medical image;

[0026] S4.2: Based on the above subscripts, perform a mixed operation on the grayscale of the voxel points of these subscripts. The mixed operation includes addition and subtraction operations, plus a larger bias (the bias value can be given independently and is not a fixed value. The user can set it according to needs. It can be changed to adding a positive bias as a variable, and adding a negative bias as suppression) to brighten the grayscale of this type of substance, subtract the bias to achieve black blood pressure suppression, and / or multiply by a scale parameter to amplify the grayscale of this part of the substance.

[0027] In the above technical solution, in S4.2, the mixed operation also includes multiplication and division operations;

[0028] The grayscale of these subscripted voxel points is multiplied by a scale parameter to amplify the grayscale of this part of the material.

[0029] Dividing the grayscale of these subscripted voxel points by a scale parameter can suppress the grayscale of the substance.

[0030] The formula is as follows: y = (x-bias) * scale + bias

[0031] Among them, y is the new grayscale and x is the original grayscale.

[0032] Is CTA the same as CT (Computed tomography angiography) vascular imaging?

[0033] The * above represents the mathematical symbol multiplication sign.

[0034] The black blood software of the present invention includes an image selection module, an ROI file selection module, a slicing module (including modules for selecting slicing modes such as coronal, sagittal, and transverse sections, a slicing position selection module, and an incremental layer number selection module), a suppression parameter control module, a calculation and mapping module, an image display module, and a suppression result storage module. The image selection module is used to select two configured images of different energy levels for import; and to select an image at one of the energy levels as the final reference image.

[0035] The ROI file selection module is used to select the imported ROI file;

[0036] The slicing module is used to select the target slicing method, slicing position and then increase the number of layers;

[0037] The suppression parameter control module is used to adjust the black blood suppression parameters of the target, highlight the cardiac cavity wall or tube wall by suppressing the blood signal, calculate and form the black blood image, and adjust the black blood target image display parameters.

[0038] The method for using the Black Blood software of the present invention comprises the following steps:

[0039] Step S51: Find the file location, double-click the application, and open the Black Blood software;

[0040] Step S52: In the Black Blood software interface, click on the “Image Selection Module” to import two configured images of different energy levels;

[0041] Step S53: Click the “reference image selection module” to select an image of any energy level in the previous step (step S52) as the final reference image;

[0042] Step S54: selecting the imported ROI file in the ROI file selection module;

[0043] Step S55: After the above image and ROI file are imported, the slicing mode is turned on;

[0044] Step S56: The main interface displays the black blood image processed in the above steps;

[0045] Step S57: The image position is adjusted by "slice position", the image grayscale is adjusted by "window width, window level", and the grayscale suppression degree is adjusted by "suppression parameter".

[0046] The principle of the CTA black blood imaging method based on substance separation of the present invention is:

[0047] Taking advantage of the different grayscale spatial distribution characteristics of different tissues in the human body, the blood flow intensity in the vascular cavity is suppressed. The blood in the resulting image appears black and low-density, while the vascular wall appears high-density, achieving the purpose of "separating" the blood in the lumen and the vascular wall. This allows the naked eye to better assess the condition of the vascular wall and prevent the missed detection or misdetection of minor vascular wall lesions.

[0048] The substances here refer to tissues with different grayscale distributions in the human body, specifically blood and blood vessel walls.

[0049] The present invention has the following advantages:

[0050] The present invention has pioneered a new method for CTA-based black blood imaging. By suppressing blood signals, it highlights the walls of the cardiac chambers or vessels, providing a high level of intracavitary anatomical details and better assessing lesions in the cardiac walls and arterial walls. This method not only overcomes the shortcomings of current magnetic resonance black blood imaging, but also significantly reduces examination costs and shortens examination time (the present invention is CT post-processing software. After a CT examination, the image is imported into the post-processing software of the present invention for processing to obtain a black blood image). The scanning time of the CT examination used in the present invention is shorter than that of MR (CT examination takes less than 1 minute per examination, while MR examination takes more than half an hour per examination), and the examination cost is also lower than MR (the scanning cost of the CT examination used in the present invention is approximately 1,000 yuan lower per examination than that of MR), thus having high clinical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 Schematic diagram of the two-dimensional grayscale space of the present invention.

[0052] Figure 2 It is a process flow chart of the present invention.

[0053] Figure 3 This is the abdominal aorta black blood imaging (coronary position) in Example 1 of the present invention. Figure 3 A in the middle is the original CTA image. Figure 3 middle B is the "black blood" image after compression.

[0054] Figure 4 This is an image of the aortic wall in Example 2 of the present invention.

[0055] Figure 5 It is the folder of the black blood software of the present invention.

[0056] Figure 6 This is a screenshot of the Black Blood software interface of the present invention.

[0057] Figure 7 This is a window diagram for selecting two configured images of different energy levels to import when operating the Black Blood software of the present invention.

[0058] Figure 8 This is a partial diagram of the operating interface for selecting an image at one of the energy levels as the final reference image when operating the Black Blood software of the present invention.

[0059] Figure 9 This is a partial diagram of the operation interface for importing ROI files when operating the Black Blood software of the present invention.

[0060] Figure 10 This is a partial diagram of the operating interface for opening the slicing module when operating the Black Blood software of the present invention.

[0061] Figure 11 This is an operational flow chart of the Black Blood software of the present invention.

[0062] exist Figure 1 In the figure, the horizontal axis represents the distribution of the two components in the gray direction at the first energy level; the vertical axis represents the distribution of the two components in the gray direction at the second energy level; Figure 1 It can be seen that in the original image, the density of the blood vessel wall and blood is close and difficult to distinguish, which is not conducive to the doctor's assessment of subtle lesions in the blood vessel wall (that is, when viewed from any energy level alone, the grayscale distribution of the two components is relatively continuous, which is not conducive to distinguishing the two (that is, two components at the same energy level)). Because the present invention suppresses the blood signal, the blood vessel wall is also highlighted (the black blood image can be obtained by suppressing the blood in the blood vessel cavity, leaving only the image of the blood vessel wall). Therefore, the method of the present invention can well distinguish the two components at the same energy level.

[0063] exist Figure 3 In the figure, A is the original image of CTA imaging, showing the abdominal aorta, Figure 3 As can be seen in A, during angiography, the vessel wall and the blood in the lumen are enhanced simultaneously. The vessel wall and the blood have the same density, which poorly displays the vessel wall and limits the ability to evaluate the vessel wall. It is impossible to evaluate the condition of the vessel wall alone. B is a black blood image formed after processing the original image in 3A using the method of the present invention. Figure 3 In B, we can see: Figure 3 The blood vessel density (blood) in image A has been suppressed, leaving only the blood vessel walls. The blood vessel walls are highlighted, which helps doctors better evaluate blood vessel wall lesions.

[0064] Figure 3 The middle arrow points to the target artery.

[0065] Figure 4 The middle arrow points to the target artery. DETAILED DESCRIPTION

[0066] The following detailed description of the embodiments of the present invention is given in conjunction with the accompanying drawings, which do not limit the present invention but are merely examples. The description makes the advantages of the present invention clearer and easier to understand.

[0067] The present invention first adopts Gaussian mixture model (GMM, Figure 1) distinguishes the n-dimensional grayscale space of material distribution, and refers to the grayscale information of the material in the ROI file area to determine its category, and finally calculates the probability that each point in the grayscale space belongs to that category. Then, by setting an empirical threshold alpha (for example, 0.95), the information in the real space of the points in the grayscale space that have a probability of belonging to the same category of ROI greater than alpha is retained. The spatial information obtained from the medical image that needs to suppress the blood components is processed (such as scaling and translation) on the grayscale of the voxel points in these spaces, so that the grayscale of the voxels close to the components of the target ROI is highlighted or darkened, thereby highlighting the statistically significant tissue boundary information.

[0068] As shown in the accompanying figures, a method for CTA black blood imaging based on material separation first delineates a region of interest (ROI) within the subject's blood vessels to extract blood information. Secondly, using the grayscale spatial distribution of two images with different energy levels and the ROI, components that are not statistically different from the ROI region are suppressed to obtain a corresponding, targeted, suppressed black blood image, highlighting the vessel wall.

[0069] The specific method includes the following steps:

[0070] Step 1: Acquire several enhanced CT images of human blood vessels;

[0071] Obtain n (n ≥ 1) spatially registered reference medical images (the following uses the common spectral CTA two-level image as an example) to form an n-dimensional grayscale space of material distribution;

[0072] Step 2: Draw the region of interest (ROI) in the blood vessel and extract blood information;

[0073] Step 3: For two images and ROIs with different energy levels, use the grayscale spatial distribution to find the components that have no statistical difference with the ROI area and suppress them;

[0074] Step 4: Obtain the corresponding target suppressed black blood image to highlight the condition of the blood vessel wall. There are different grayscale spatial distributions on the DICOM (a special image format for medical images) reconstructed image. The grayscale of each pixel represents the basic information of different organs and tissues. After the contrast agent in the blood vessel forms a high CT value, the grayscale difference from the surrounding blood vessel wall is large. By obtaining the intravascular blood flow ROI information, the present invention uses the grayscale spatial distribution to find the components that have no statistical difference with the ROI area for suppression, and obtains the corresponding target suppressed black blood image, highlighting the condition of the blood vessel wall, which is more conducive to the physician's assessment of the wall atherosclerotic plaque, intramural hematoma or tumor invasion of the blood vessel wall and other conditions.

[0075] Furthermore, in step 1, the spatially registered images include a low-energy image and a high-energy image.

[0076] Furthermore, in step 1, the specific method for obtaining several images of human blood vessels enhanced by CT is:

[0077] S1.1: Use a CT machine to acquire the medical images required for processing;

[0078] S1.2: Align all reference images so that the tissue distribution of the two images corresponds pixel by pixel.

[0079] Furthermore, in step 2, a region of interest (ROI) is delineated within the target blood vessel, and an ROI file that is consistent with the reference medical image space is obtained to determine the substance to be separated.

[0080] Furthermore, in step 3, the Gaussian mixture model (GMM) method is used to distinguish the n-dimensional grayscale space of the material distribution. At the same time, the material grayscale information of the ROI file area is referred to determine its category. Finally, the probability of each point in the grayscale space belonging to the category is calculated.

[0081] The specific method to obtain the probability that each point in the gray space belongs to this class is:

[0082] S31: Empirically estimate the number of components in the medical image (for spectral CT, 64 is generally selected) as the parameter of the GMM model, that is, divide the distribution of the n-dimensional grayscale space into 64 different material components to complete material separation.

[0083] S32: Based on the existing ROI, extract the grayscale distribution in this area, and then bring it into the GMM model to get which category the material in this ROI area belongs to; use the GMM model to calculate the probability that all grayscales belong to this category, such as Figure 1 shown.

[0084] Furthermore, in step three, by setting an empirical threshold alpha (for example, 0.95), the information of points in the grayscale space that have a probability greater than alpha belonging to the same type of ROI in the real space is retained.

[0085] Furthermore, in step 4, a medical image showing the blood components to be suppressed is obtained, and the spatial information obtained in step 3 is brought into the image. The grayscale of the voxel points in these spaces is pre-processed (such as scaling and translation), so that the grayscale of the voxels close to the components of the ROI in step 2 is darkened and suppressed, thereby highlighting the statistically significant tissue boundary information, and obtaining the suppressed CTA black blood image ( Figure 3 );

[0086] The specific method is:

[0087] S4.1: After obtaining the spatial information from step 3, the spatial coordinates of the voxel point of this type are obtained, i.e., the index subscript of the grayscale matrix of the medical image;

[0088] S4.2: Based on the above subscripts, a mixing operation is performed on the grayscale of the voxel points of these subscripts. The mixing operation includes addition and subtraction operations. A larger bias is added to brighten the grayscale of this type of substance, and the bias is subtracted to achieve black blood pressure control, and / or a scale parameter can be multiplied to amplify the grayscale of this part of the substance.

[0089] Furthermore, in S4.2, mixed operations also include multiplication and division operations;

[0090] The grayscale of these subscripted voxel points is multiplied by a scale parameter to amplify the grayscale of this part of the material.

[0091] Example 1

[0092] This embodiment uses the method of the present invention to perform black blood imaging on the abdominal aorta (coronary position) of a normal volunteer. The specific method is as follows:

[0093] Step 1: If Figure 5 、 Figure 6 As shown, open the Black Blood software and obtain several enhanced CT images of the abdominal aorta of a normal volunteer (such as Figure 3 As shown in Figure A, it can be seen that during angiography, the vascular wall and the blood in the lumen are enhanced simultaneously. The vascular wall and the blood have the same density, which poorly displays the vascular wall and limits the ability to assess the vascular wall. It is impossible to assess the condition of the vascular wall alone).

[0094] Acquire two spatially registered reference medical images (referring to spectral CT angiography images) to construct a 2D grayscale space of material distribution and obtain the scanned original CTA image;

[0095] S1.1: Use a spectral CT machine to acquire enhanced medical images. The enhanced medical image can be one;

[0096] S1.2: All reference images (i.e., the two spatially registered reference medical images obtained in step 1) are registered so that the tissue distributions of the two (i.e., images of different energy levels) correspond pixel by pixel.

[0097] Step 2: Delineate a region of interest (ROI) within the blood vessels of the image obtained in step 1, obtain an ROI file that is spatially consistent with the reference medical image (i.e., the reference medical image in step 1), determine the substance to be separated, and extract the blood information on the image obtained in step 1;

[0098] Step 3: If Figure 7As shown, first select two configured images of different energy levels to import (in the test data file), then select one of the energy level images as the final reference image, as shown Figure 8 As shown; Finally, import the ROI file, such as Figure 9 As shown;

[0099] For two images and ROIs with different energy levels, the grayscale spatial distribution is used to find the components that have no statistical difference with the ROI area and suppress them;

[0100] S31: Empirically estimate the number of components in the medical image as a parameter of the GMM model;

[0101] S32: Based on the existing ROI, extract the grayscale distribution within this area, and then bring it into the GMM model to determine which category the material in this ROI area belongs to; use the GMM model to calculate the probability that all grayscales belong to this category;

[0102] By setting the empirical threshold alpha, the information of points in the grayscale space whose probability of belonging to the same ROI is greater than alpha in the real space is retained.

[0103] Step 4: If Figure 10 As shown, turn on the slice module and enter the slice observation mode to obtain the black blood image of the target blood vessel, as shown in Figure 3 As shown in B, the black blood image corresponding to the target is obtained, highlighting the condition of the blood vessel wall.

[0104] S4.1: After obtaining the spatial information from step 3, the spatial coordinates of the voxel point of this type are obtained, i.e., the index subscript of the grayscale matrix of the medical image;

[0105] S4.2: Based on the above subscripts, a mixing operation is performed on the grayscale of the voxel points with these subscripts. The mixing operation includes addition and subtraction operations. Adding a larger bias brightens the grayscale of this type of substance, and subtracting the bias achieves black blood pressure control.

[0106] Conclusion: In this embodiment, the black blood image (coronal position) obtained by suppressing the original CTA image using the method of the present invention is as follows: Figure 3 As shown in B, from Figure 3 In B, we can see: Figure 3 The blood vessel density (blood) in image A has been suppressed, leaving only the vessel wall. The vessel wall is highlighted, which helps doctors better assess vessel wall lesions. As can be seen, the black blood image obtained by the method of the present invention in this embodiment can better display the normal condition of the arterial wall.

[0107] Example 2

[0108] This embodiment adopts the method of the present invention to perform black blood imaging on the thoracic aorta (transverse axis) of a patient with intramural hematoma. The specific method is the same as that of Example 1. The difference is that: in step 1, several CT enhanced images of the thoracic aorta of the patient with intramural hematoma are obtained (such as Figure 4 As shown in the left figure, during angiography, the density of the vessel wall and lumen is high, and intramural hematoma lesions are poorly displayed, making it easy for doctors to miss the diagnosis);

[0109] This embodiment acquires two sets of spatially registered medical images (referring to spectral CT angiography images) to form a three-dimensional grayscale space of material distribution.

[0110] Conclusion: In this embodiment, the black blood image (coronal position) obtained by suppressing the original CTA image using the method of the present invention is as follows: Figure 4 As shown in the right figure, from Figure 4 The right picture shows: Figure 4 In the left image, the blood within the lumen has been suppressed, leaving only the vessel wall visible. This allows for better visualization of the vessel wall and intramural hematoma lesions, facilitating lesion detection and helping physicians better assess vessel wall lesions. As can be seen, the black blood image obtained using the method of the present invention in this embodiment can better visualize intramural hematoma lesions, preventing the inability to distinguish between the inner and outer vessel walls due to the same or similar density between the vessel wall and lumen, resulting in a lack of accurate vessel wall information and the potential for missed or misdiagnosed minor lesions.

[0111] It should be noted that the specific embodiments described herein are merely illustrative of the spirit of the present invention. Persons skilled in the art may make various modifications, additions, or substitute similar methods (e.g., using the k-nearest neighbor (KNN) method as an alternative to substance identification) to the described specific embodiments without departing from the spirit of the present invention or exceeding the scope of the appended claims.

[0112] Other parts not described belong to the prior art.

Claims

1. A method for CTA black blood imaging based on substance separation, characterized by: The following steps are included: Step 1: Acquire several enhanced CT images of human blood vessels; Obtain n spatially registered reference medical images, where n ≥ 1, constituting an n-dimensional grayscale space of material distribution; Step 2: Draw a region of interest (ROI) within the blood vessels of the image in step 1 and extract blood information from the image; Step 3: For two images and ROIs with different energy levels, use the grayscale spatial distribution to find the components that have no statistical difference with the ROI area and suppress them; Step 4: Obtain the black blood image corresponding to the target suppression to highlight the condition of the blood vessel wall; In step 1, the specific method for obtaining several images of human blood vessels enhanced by CT is as follows: S1.1: Use spectral or dual-energy CT to acquire enhanced medical images; S1.2: Register all reference medical images so that the tissue distribution of the two registered images corresponds pixel by pixel; In step 3, the Gaussian mixture model method is used to distinguish the n-dimensional grayscale space of material distribution, and the material grayscale information of the ROI file area is referred to determine its category. Finally, the probability of each point in the grayscale space belonging to the category is calculated; S3.1: Empirically estimate the number of components in the medical image as parameters of the GMM model; S3.2: Based on the existing ROI, extract the grayscale distribution within this area, and then use the GMM model to determine which category the material in this ROI area belongs to; use the GMM model to calculate the probability that all grayscale values ​​belong to this category; In step 4, a medical image displaying the blood components to be suppressed is obtained and the spatial information obtained in step 3 is brought into play. The grayscale of the voxel points in these spaces is processed so that the grayscale of the voxels close to the components of the ROI in step 2 is darkened and suppressed, thereby highlighting statistically significant tissue boundary information and obtaining a suppressed CTA black blood image. In step 4, the specific method is: S4.1: After obtaining the spatial information from step 3, the spatial coordinates of the voxel point of this type are obtained, i.e., the index subscript of the grayscale matrix of the medical image; S4.2: Based on the above subscripts, perform a blending operation on the grayscale of the voxel points with these subscripts. The blending operation includes addition and subtraction operations. Adding a larger bias will brighten the grayscale of this type of substance, and subtracting the bias will achieve black blood pressure modulation. In S4.2, mixed operations also include multiplication and division; Multiply the grayscale of these subscripted voxel points by a scale parameter to amplify the grayscale of this part of the material; The formula is as follows: y = (x-bias) * scale + bias Among them, y is the new grayscale and x is the original grayscale.

2. The method for CTA black blood imaging based on substance separation according to claim 1, characterized in that: In step one, the spatially registered images include images at different energy levels.

3. The method for CTA black blood imaging based on substance separation according to claim 2, characterized in that: In step 2, a region of interest (ROI) is outlined within the target blood vessel, and an ROI file that is consistent with the reference medical image space is obtained to determine the substance to be separated.

4. The method for CTA black blood imaging based on substance separation according to claim 3, characterized in that: In step three, by setting an empirical threshold alpha, the information of points in the grayscale space that belong to the same type of ROI with a probability greater than alpha in the real space is retained.

Images