Component analysis method, device, computer equipment and storage medium

Abstract

The present application relates to a component analysis method, device, computer equipment and storage medium. The method comprises: for any target image in a multi-series image, according to the vascular centerline and target vascular plaque of the target image, N vascular cross-sectional images are created on N levels, wherein N is an integer greater than 0; according to the target vascular cross-sectional images of each target image on the i-th level, the component probability image of the target vascular plaque on the i-th level is determined, wherein i is an integer greater than 0 and less than or equal to N; according to the component probability image and the target vascular cross-sectional images of each target image on the i-th level, the component distribution image corresponding to the target vascular plaque on the i-th level is obtained. The use of this method can improve the accuracy of component analysis.

Description

Technical Field

[0001] The present application relates to the field of biomedical technology, and in particular to a component analysis method, device, computer equipment and storage medium. Background Art

[0002] Magnetic resonance imaging is a type of tomographic imaging. It uses the magnetic resonance phenomenon to obtain electromagnetic signals from the human body and reconstruct human body information. It is widely used in the medical field.

[0003] In the related art, after obtaining multiple sequence images through magnetic resonance imaging, the user needs to observe the images with the naked eye to assist in diagnosis and treatment based on the observation results. For example, in the analysis of vascular plaque components, after the user determines the suspicious area in the sequence image, the user compares the signal level of the suspicious area with the surrounding sternocleidomastoid muscle area with the naked eye, and then obtains the component analysis result based on the signal level observed by the naked eye. With this analysis method, the component analysis result is not objective and has low accuracy. Summary of the invention

[0004] Based on this, it is necessary to provide a component analysis method, device, computer equipment and computer-readable storage medium that can improve the accuracy of vascular plaque component analysis in response to the above technical problems.

[0005] In a first aspect, the present application provides a component analysis method, the method comprising:

[0006] For any target image in the multiple sequence images, create N vascular cross-sectional images on N levels according to the vascular centerline and target vascular plaque of the target image, where N is an integer greater than 0;

[0007] Determine a component probability image of the target vascular plaque at the i-th level according to the target vascular cross-sectional image at the i-th level of each of the target images, wherein i is an integer greater than 0 and less than or equal to N;

[0008] A component distribution image corresponding to the target blood vessel plaque at the i-th level is obtained according to the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level.

[0009] In one embodiment, obtaining a component distribution image corresponding to the target blood vessel plaque at the i-th level according to the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level includes:

[0010] Obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image and the vascular cross-sectional image of each target image at the i-th level;

[0011] The blood vessel cross-sectional image includes at least one of the target blood vessel cross-sectional image and a histogram-equalized image obtained by performing histogram-equalization processing on the target blood vessel cross-sectional image.

[0012] In one embodiment, obtaining the component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image and the vascular cross-sectional image of each target image at the i-th level includes:

[0013] Obtaining a correlation analysis image corresponding to the target blood vessel plaque according to each of the target images and / or the target blood vessel cross-sectional image of each of the target images at the i-th level;

[0014] Obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image, the vascular cross-sectional image of each target image at the i-th level, and the correlation analysis image;

[0015] The correlation analysis image includes at least one of a distance field image corresponding to the target blood vessel cross-sectional image and an intensity contrast image corresponding to each of the target blood vessel cross-sectional images.

[0016] In one embodiment, determining the component probability image of the target blood vessel plaque at the i-th level according to the target blood vessel cross-sectional image at the i-th level of each target image includes:

[0017] Perform segmentation processing according to the multiple sequence images to obtain a region of interest ROI image corresponding to the multiple sequence images;

[0018] Obtaining statistical images corresponding to each of the target images according to the ROI image and the corresponding area of ​​the ROI image in each of the target images;

[0019] For any of the target images, an intensity contrast image is obtained according to the target blood vessel cross-sectional image on the i-th layer and the statistical image corresponding to the target image, wherein the intensity contrast image is used to characterize the signal intensity contrast result between the target blood vessel cross-sectional image and the statistical image;

[0020] According to the intensity contrast image and component distribution prior information corresponding to each of the target images, a component probability image of the target vascular plaque at the i-th level is obtained.

[0021] In one embodiment, the method further comprises:

[0022] According to a cross-sectional image of a target blood vessel at an i-th level in at least one of the target images, a lumen and wall segmentation image of the multiple sequence images at the i-th level is obtained;

[0023] Determine a first distance between each pixel point inside the tube wall and the tube lumen contour in the tube lumen and tube wall segmentation image, and a second distance between each pixel point and the tube wall contour;

[0024] According to the first distance and the second distance corresponding to each pixel point, a distance field image on the i-th layer is obtained, and the pixel value of any pixel point in the distance field image is used to characterize the distance between the pixel point and the lumen contour and the tube wall contour.

[0025] In one embodiment, the method further comprises:

[0026] Statistical analysis is performed on the component distribution images at N levels to obtain component analysis results of the target vascular plaque.

[0027] In one embodiment, the method implements the component analysis network according to the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level to obtain the component distribution image corresponding to the target blood vessel plaque at the i-th level, and the method further includes:

[0028] According to a preset first training set, the analysis network is trained to obtain the component analysis network;

[0029] The first training set includes multiple first sample groups, which include multiple sequence sample images, N layers of sample blood vessel cross-sectional images corresponding to the multiple sequence sample images, and annotated component distribution images of each layer of the sample blood vessel cross-sectional images.

[0030] In one embodiment, the method implements the segmentation processing according to the multiple sequence images through a segmentation network to obtain a region of interest ROI image corresponding to the multiple sequence images, and the method further includes:

[0031] Training the initial segmentation network according to a preset second training set to obtain the segmentation network;

[0032] The second training set includes a plurality of second sample groups, and the second sample groups include multiple sequence sample images and annotated ROI images corresponding to the multiple sequence sample images.

[0033] In a second aspect, the present application also provides a component analysis device. The device comprises:

[0034] A creation module, for creating N vascular cross-sectional images on N levels according to the vascular centerline and target vascular plaque of any target image in the multiple sequence images, wherein N is an integer greater than 0;

[0035] A first determination module, configured to determine a component probability image of the target vascular plaque at the i-th level according to a target vascular cross-sectional image at the i-th level of each of the target images, wherein i is an integer greater than 0 and less than or equal to N;

[0036] The first processing module is used to obtain a component distribution image corresponding to the target blood vessel plaque at the i-th level according to the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level.

[0037] In one embodiment, the first processing module is further used to:

[0038] Obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image and the vascular cross-sectional image of each target image at the i-th level;

[0039] The blood vessel cross-sectional image includes at least one of the target blood vessel cross-sectional image and a histogram-equalized image obtained by performing histogram-equalization processing on the target blood vessel cross-sectional image.

[0040] In one embodiment, the first processing module is further used to:

[0041] Obtaining a correlation analysis image corresponding to the target blood vessel plaque according to each of the target images and / or the target blood vessel cross-sectional image of each of the target images at the i-th level;

[0042] Obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image, the vascular cross-sectional image of each target image at the i-th level, and the correlation analysis image;

[0043] The correlation analysis image includes at least one of a distance field image corresponding to the target blood vessel cross-sectional image and an intensity contrast image corresponding to each of the target blood vessel cross-sectional images.

[0044] In one embodiment, the first determining module is further used to:

[0045] Perform segmentation processing according to the multiple sequence images to obtain a region of interest ROI image corresponding to the multiple sequence images;

[0046] Obtaining statistical images corresponding to each of the target images according to the ROI image and the corresponding area of ​​the ROI image in each of the target images;

[0047] For any of the target images, an intensity contrast image is obtained according to the target blood vessel cross-sectional image on the i-th layer and the statistical image corresponding to the target image, wherein the intensity contrast image is used to characterize the signal intensity contrast result between the target blood vessel cross-sectional image and the statistical image;

[0048] According to the intensity contrast image and component distribution prior information corresponding to each of the target images, a component probability image of the target vascular plaque at the i-th level is obtained.

[0049] In one embodiment, the device further comprises:

[0050] A second processing module is used to obtain a lumen and wall segmentation image of the multiple sequence images at the i-th level according to a cross-sectional image of a target blood vessel at the i-th level in at least one of the target images;

[0051] A second determination module is used to determine a first distance between each pixel point inside the tube wall and the tube lumen contour in the tube lumen and tube wall segmentation image, and a second distance between each pixel point and the tube wall contour;

[0052] The third processing module is used to obtain the distance field image on the i-th layer according to the first distance and the second distance corresponding to each pixel point, and the pixel value of any pixel point in the distance field image is used to represent the distance between the pixel point and the lumen contour and the tube wall contour.

[0053] In one embodiment, the device further comprises:

[0054] The fourth processing module is used to perform statistical analysis based on the component distribution images at N levels to obtain component analysis results of the target vascular plaque.

[0055] In one embodiment, the device further comprises:

[0056] A first training module, used for training the analysis network according to a preset first training set to obtain the component analysis network;

[0057] The first training set includes multiple first sample groups, which include multiple sequence sample images, N layers of sample blood vessel cross-sectional images corresponding to the multiple sequence sample images, and annotated component distribution images of each layer of the sample blood vessel cross-sectional images.

[0058] In one embodiment, the device further comprises:

[0059] A second training module, used for training the initial segmentation network according to a preset second training set to obtain the segmentation network;

[0060] The second training set includes a plurality of second sample groups, and the second sample groups include multiple sequence sample images and annotated ROI images corresponding to the multiple sequence sample images.

[0061] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the above component analysis method when executing the computer program.

[0062] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the above component analysis method when executed by a processor.

[0063] In a fifth aspect, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the above component analysis method is implemented.

[0064] The component analysis method, device, computer equipment and storage medium provided in the embodiment of the present application, for any target image in a multi-sequence image, creates N cross-sectional images of blood vessels on N levels according to the blood vessel centerline and target blood vessel plaque of the target image, wherein N is an integer greater than 0; according to the cross-sectional images of the target blood vessels of each target image at the i-th level, determines the component probability image of the target blood vessel plaque at the i-th level; according to the component probability image and the cross-sectional images of the target blood vessels of each target image at the i-th level, obtains the component distribution image corresponding to the target blood vessel plaque at the i-th level. Based on the component analysis method, device, computer equipment and storage medium provided in the embodiment of the present application, the component probability image of the target blood vessel plaque can be obtained by analysis, and then the component probability image can be automatically and objectively analyzed to obtain the component distribution image of the target blood vessel plaque, which can avoid the influence of subjectivity when performing component analysis by human eyes, making the component analysis result more objective and accurate, and improving the accuracy and efficiency of component analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 Schematic diagram of a component analysis method in one embodiment;

[0066] Figure 2 Schematic diagram of a component analysis method in one embodiment;

[0067] Figure 3 Schematic diagram of a component analysis method in one embodiment;

[0068] Figure 4 Schematic diagram of a component analysis method in one embodiment;

[0069] Figure 5 Schematic diagram of a component analysis method in one embodiment;

[0070] Figure 6a to Figure 6b Schematic diagram of a component analysis method in one embodiment;

[0071] Figure 7 is a structural block diagram of a component analysis device in one embodiment;

[0072] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0073] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0074] In one embodiment, Figure 1 As shown, a component analysis method is provided. This embodiment uses the method applied to a terminal as an example for illustration. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0075] Step 102, for any target image in the multiple sequence images, create N vascular cross-sectional images on N levels according to the vascular centerline and target vascular plaque of the target image, where N is an integer greater than 0.

[0076] In an embodiment of the present application, any sequence image in the multiple sequence images is a multi-contrast magnetic resonance image, which may include: multiple sequence images (hereinafter referred to as images) collected for the same target object through multiple imaging technologies such as T1 weighted imaging T1WI, T2 weighted imaging T2WI, TOF (time-of-flight, based on the time flight principle) imaging, enhanced T1WI and other imaging technologies.

[0077] The multiple sequence images in the embodiment of the present application may be multiple images that have been spatially aligned after the registration operation. For example, after multiple images are acquired using multiple imaging technologies, one or a series of spatial transformations may be performed to make the corresponding points in each image spatially consistent, that is, the same part of the same target object has the same spatial position in multiple images.

[0078] The blood vessel centerline in the multi-series image can be created based on one or more target images in the multi-series image (for example, refer to Figure 2 As shown) and target vascular plaques. Exemplarily, a vascular centerline extraction operation can be performed on one or more target images through a pre-trained neural network for extracting vascular centerlines to obtain vascular centerlines in multi-sequence images, or the vascular centerlines in multi-sequence images can be marked by a user through manual operation in a terminal interface. Similarly, vascular plaque extraction can be performed on one or more target images through a pre-trained neural network for extracting vascular plaques to obtain vascular plaques in multi-sequence images, and the target vascular plaque can be any vascular plaque among the detected vascular plaques. Alternatively, the user can determine the starting and ending points of the vascular plaque in the terminal interface and then depict the target vascular plaque. The above is only an example of extracting vascular centerlines and target vascular plaques, and the embodiments of the present application do not provide specific examples of the methods of extracting vascular centerlines and target vascular plaques.

[0079] After obtaining the vascular centerline and the target vascular plaque, for any target image, N points can be determined on the area corresponding to the target vascular plaque in the vascular centerline (hereinafter referred to as the truncation area), and each target image can be reconstructed using the N points. For any target image, N vascular cross-sectional images can be obtained.

[0080] For example, the center point P0 of the truncated area can be determined, and the center point P0 is used as the midpoint, and k points are obtained in the front and rear directions of the blood vessel with a step length of λ (λ can be a preset value), that is, (2k+1) points are taken from the truncated area, and the corresponding blood vessel cross-sectional images are reconstructed with each point as the center. For example, taking point P0 as an example, the blood vessel cross-sectional image can be reconstructed with the XY axis of P0 as the direction, and by analogy, (2k+1) blood vessel cross-sectional images can be obtained.

[0081] Step 104, determining a component probability image of the target blood vessel plaque at the i-th level based on the target blood vessel cross-sectional image of each target image at the i-th level, wherein i is an integer greater than 0 and less than or equal to N.

[0082] In the embodiment of the present application, the cross-sectional image of the blood vessel of each target image at the i-th level can be determined as the cross-sectional image of the target blood vessel. The component probability image of the target blood vessel plaque at the i-th level can be determined by the cross-sectional image of the target blood vessel of each target image in the multi-series image, and the component probability image can be used to represent the probability of each position in the target blood vessel plaque at the i-th level being the component to be detected.

[0083] Exemplarily, the component probability image of the target vascular plaque at the i-th level can be obtained by statistically analyzing the target vascular cross-sectional images of each target image. For example, the target vascular cross-sectional images of each target image can be statistically analyzed by a pre-trained neural network for generating component probability images to obtain the component probability image of the target vascular plaque at the i-th level; or, a reference area of ​​the vascular plaque (for example, the sternocleidomastoid muscle area) can be obtained, and the target vascular cross-sectional images of each target image can be compared and analyzed with the signal height in the reference area, and the probability of each pixel being the component to be detected is determined based on the comparison and analysis results, thereby obtaining the corresponding component probability image. The embodiment of the present application does not specifically limit the generation method of the component probability image.

[0084] Step 106, obtaining a component distribution image corresponding to the target blood vessel plaque at the i-th level based on the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level.

[0085] In an embodiment of the present application, component analysis processing can be performed on the component probability image and the target cross-sectional image of each target image at the i-th level to obtain a component distribution image corresponding to the target vascular plaque at the i-th level. The component distribution image is used to represent the components corresponding to each position in the target vascular plaque. For example, different colors can be used to represent different components in the component distribution image. For example, if the component includes at least one of lipid necrotic core, calcification, hemorrhage, loose stroma, fibrous cap rupture-niche shadow, etc., then different components can be marked with different colors in the component distribution image.

[0086] For example, in the embodiment of the present application, a component analysis process can be performed on the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level through a pre-trained neural network to obtain a component distribution image corresponding to the target blood vessel plaque at the i-th level. Similarly, component distribution images corresponding to the target blood vessel plaque at N levels can be obtained.

[0087] A component analysis method provided in an embodiment of the present application creates N vascular cross-sectional images on N levels according to the vascular centerline and target vascular plaque of the target image for any target image in a multi-sequence image; determines the component probability image of the target vascular plaque on the i-th level according to the target vascular cross-sectional image of each target image on the i-th level; obtains the component distribution image corresponding to the target vascular plaque on the i-th level according to the component probability image and the target vascular cross-sectional image of each target image on the i-th level. Based on the component analysis method provided in an embodiment of the present application, the component probability image of the target vascular plaque can be obtained by analysis, and then the component probability image can be automatically and objectively analyzed to obtain the component distribution image of the target vascular plaque, which can avoid the influence of subjectivity when performing component analysis by human eyes, making the component analysis result more objective and accurate, and improving the accuracy and efficiency of component analysis.

[0088] In one embodiment, in step 106, a component distribution image corresponding to the target vascular plaque at the i-th level is obtained based on the component probability image and the target vascular cross-sectional image of each target image at the i-th level, including:

[0089] According to the component probability image and the blood vessel cross-sectional image of each target image at the i-th level, a component distribution image corresponding to the target blood vessel plaque at the i-th level is obtained;

[0090] The blood vessel cross-sectional image includes at least one of a target blood vessel cross-sectional image and a histogram-equalized image obtained by performing histogram-equalization processing on the target blood vessel cross-sectional image.

[0091] In an embodiment of the present application, histogram equalization processing can be performed on each target blood vessel cross-sectional image to obtain a histogram equalized image corresponding to each target blood vessel cross-sectional image, and component analysis processing can be performed based on the target blood vessel cross-sectional image and / or histogram equalized image at the i-th level of each target image, as well as the component probability image, to obtain a component distribution image corresponding to the target blood vessel plaque at the i-th level.

[0092] Based on the component analysis method provided in the embodiment of the present application, since the histogram-equalized image can enhance the image contrast, combining the histogram-equalized image with the component analysis can further improve the accuracy of the component analysis, thereby obtaining a higher-precision component distribution image.

[0093] In one embodiment, referring to Figure 3 , according to the component probability image and the vascular cross-sectional image of each target image at the i-th level, the component distribution image corresponding to the target vascular plaque at the i-th level is obtained, including:

[0094] Step 302, obtaining a correlation analysis image corresponding to the target blood vessel plaque according to each target image and / or the target blood vessel cross-sectional image of each target image at the i-th layer;

[0095] Step 304, obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image, the vascular cross-sectional image of each target image at the i-th level, and the correlation analysis image;

[0096] The correlation analysis image includes at least one of a distance field image corresponding to the target blood vessel cross-sectional image and an intensity contrast image corresponding to each target blood vessel cross-sectional image.

[0097] In the embodiment of the present application, a segmented image of the lumen and wall of the blood vessel can be obtained by analyzing the cross-sectional image of the target blood vessel at the i-th level of any target image, and a distance field image can be obtained based on the distance from each pixel point to the lumen contour and the wall contour, that is, the distance field image can be used to represent the distance from each pixel point to the lumen and the wall.

[0098] In an embodiment of the present application, a reference region of a target vascular plaque can be obtained through each target image, and the reference region can be a ROI (region of interest). Exemplarily, each target image can be segmented by a pre-trained neural network for generating an ROI image to obtain an ROI image. Then, each target vascular cross-sectional image can be compared with the ROI image for signal height, and an intensity contrast image corresponding to each target vascular cross-sectional image can be obtained, that is, the intensity contrast image can be used to characterize the signal intensity contrast result between the target vascular cross-sectional image and the ROI image.

[0099] After obtaining the intensity contrast image and / or the distance field image, component analysis processing can be performed based on the intensity contrast image and / or the distance field image, the component probability image, and the vascular cross-sectional image of each target image at the i-th level to obtain a component distribution image corresponding to the target vascular plaque at the i-th level.

[0100] The component analysis method provided in the embodiment of the present application, combined with the distance field image for component analysis, can add position prior to the component analysis process, and can assist in component segmentation (for example: the fibrous cap is an area close to the vascular lumen, and the distance field image can be used to obtain the distance of each position from the vascular lumen, so as to determine to a certain extent whether it belongs to the fibrous cap component based on the distance). The component analysis of the collective intensity contrast image can simulate the user's naked eye comparison process and avoid the subjective differences of the user's naked eye comparison. Therefore, the use of distance field images and / or intensity contrast images for component analysis can further improve the accuracy of component analysis to obtain a higher accuracy component distribution image.

[0101] In one embodiment, referring to Figure 4 As shown, step 104, determining a component probability image of a target blood vessel plaque at the i-th level according to a target blood vessel cross-sectional image at the i-th level of the target image, includes:

[0102] Step 402, performing segmentation processing according to the multiple sequence images to obtain a region of interest ROI image corresponding to the multiple sequence images;

[0103] Step 404, obtaining a statistical image corresponding to each target image according to the ROI image and the corresponding area of ​​the ROI image in each target image;

[0104] Step 406, for any of the target images, an intensity contrast image is obtained according to the target blood vessel cross-sectional image on the i-th layer and the statistical image corresponding to the target image, wherein the intensity contrast image is used to represent the signal intensity contrast result between the target blood vessel cross-sectional image and the statistical image;

[0105] Step 408, obtaining a component probability image of the target vascular plaque at the i-th level according to the intensity contrast image and component distribution prior information corresponding to each target image.

[0106] In an embodiment of the present application, a multi-sequence image can be segmented using a pre-trained neural network for generating ROI images to obtain an ROI image corresponding to the multi-sequence image. The ROI image may include a reference area for comparison with the vascular plaque area, such as the sternocleidomastoid muscle area.

[0107] After the ROI image is obtained, statistical processing is performed on the ROI image and the signals in the region corresponding to the ROI image in each target image (hereinafter referred to as the reference region) to obtain the statistical image corresponding to each target image. Exemplarily, the statistical image corresponding to each target image can be obtained by performing statistical processing such as mean or median on the ROI image and the reference region.

[0108] After obtaining the statistical image corresponding to the target image, a signal comparison process can be performed between the target blood vessel cross-sectional image at the i-th level of the target image and the statistical image, and an intensity comparison image can be obtained according to the signal comparison result. Exemplarily, the signal comparison process can be a process of dividing each pixel point in the target blood vessel cross-sectional image with each pixel point in the statistical image, or a process of dividing each pixel point in the statistical image with each pixel point in the target blood vessel cross-sectional image. The embodiment of the present application takes the process of dividing the target blood vessel cross-sectional image with the statistical image as an example.

[0109] After obtaining the intensity contrast image of each target object at the i-th level, component probability analysis processing can be performed using component distribution prior information and the intensity contrast image of each target object at the i-th level to obtain a component probability image of the target vascular plaque at the i-th level.

[0110] Exemplarily, the components in the vascular plaque are mainly judged based on the adjacent muscle tissue signal as a reference standard, and the component distribution prior information can be a component prior distribution table T as shown in Table 1 below, wherein T in the component prior distribution table may include: + (high signal), 0 (equal signal), - (low signal).

[0111] Table 1

[0112] Element T1WI T2WI TOF Enhanced T1WI Lipid necrosis core 0 / + － / 0 0 － Bleeding + 0 / + + － Calcification － － － － Loose interstitial － / 0 + 0 + Fibrous cap rupture-niche － － + －

[0113] According to the component prior distribution table T, the relative distribution sequence of each component on different target images can be obtained. For example, the relative distribution sequence of calcification is (-, -, -, -). The probability that each pixel in the cross-sectional image of the target blood vessel is each component can be determined by the probability prior function P. The input information of the probability prior function P is the prior distribution table T and the intensity contrast image of each target image on the i-th layer. The output information is the probability that each pixel is each component. For any component, according to the obtained probability that each pixel is the component, the component probability image corresponding to the component can be obtained.

[0114] Exemplarily, for any intensity contrast image, it can be ternary by using a threshold. For example: for any pixel point, its pixel value can be converted into a signal "-" in the range of (0, 0.8), can be converted into a signal "0" in the range of (0.8, 1.2), and can be converted into a signal "+" in the range of (1.2, infinity). After each intensity contrast image is converted as described above, each pixel point can obtain a three-value sequence such as in the prior distribution table T. Furthermore, for any component, the distance from the three-value sequence to the relative distribution sequence corresponding to the component (for example, calcification) can be calculated. The embodiment of the present application does not specifically limit the distance calculation method. For example, distance measurement methods such as chessboard distance and Euler distance can be used.

[0115] After obtaining the distance between the three-value sequence corresponding to each pixel point and the relative distribution sequence corresponding to the component, the distance corresponding to each pixel point can be converted into a probability. For example, the distance can be converted into a probability through a probability conversion function, which can make the probability 1 when the distance is 0, and the greater the distance, the smaller the probability. The probability conversion function is called to calculate the probability of all pixel points inside the wall of the blood vessel (the area between the wall contour and the lumen contour is determined as the inside of the wall), and the probability of the pixel point outside the wall being the component is determined to be 0. The probability of all pixel points being the component constitutes the component probability image of the target vascular plaque at the i-th level.

[0116] The component analysis method provided in the embodiment of the present application has a prior distribution of the high and low characteristics of the signals of different components in the target images of different sequences and the signals in the reference area. That is, the component probability image of each component can be constructed by using the prior distribution of each component as a constraint condition, and then the component probability image can be used as a constraint condition to automatically perform component analysis, which can further improve the accuracy and efficiency of component analysis.

[0117] In one embodiment, referring to Figure 5 As shown, the component analysis method may also include:

[0118] Step 502, obtaining a lumen wall segmentation image at the i-th level of the multiple sequence images according to the target blood vessel cross-sectional image at the i-th level in at least one target image;

[0119] Step 504, determining a first distance between each pixel point inside the tube wall and the tube lumen contour, and a second distance between each pixel point and the tube wall contour in the tube lumen and tube wall segmentation image;

[0120] Step 506, obtaining a distance field image on the i-th layer according to the first distance and the second distance corresponding to each pixel point, and the pixel value of any pixel point in the distance field image is used to represent the distance between the pixel point and the lumen contour and the tube wall contour.

[0121] In an embodiment of the present application, a cross-sectional image of a target blood vessel at the i-th level can be segmented based on at least one target image to obtain a lumen and wall segmentation image at the i-th level of a multi-series image, in which the lumen area and the wall area can be identified.

[0122] For example, a pre-trained neural network for lumen and wall segmentation can be used to segment the cross-sectional image of the target blood vessel at the i-th level of at least one target image to obtain the lumen and wall segmentation image of the multiple sequence images at the i-th level. Alternatively, the user can manually draw the lumen and wall segmentation image of the multiple sequence images at the i-th level on the terminal interface.

[0123] After obtaining the lumen wall segmentation image, for each pixel point inside the wall (the area between the wall contour and the lumen contour is determined as the inside of the wall), the first distance from the lumen contour and the second distance from the wall contour can be determined. Then, based on the first distance and the second distance corresponding to each pixel point on the inner side of the wall, the distance field result is calculated to obtain the distance field result corresponding to each pixel point, and this distance field result can be used to characterize the distance of the pixel point from the lumen contour and the wall contour. Further, a distance field image can be generated according to the distance field result corresponding to each pixel point.

[0124] Exemplarily, the distance field result of each pixel point can be calculated through a distance field function. The distance field function is denoted as f(x1, x2). When x1 ≤ σ, f(x1, x2) = 1, where 0 ≤ σ < x1 + x2 and σ is a hyperparameter. When x1 > σ, f(x1, x2) = g(t), where t = |x1 - σ| / (x1 + x2); g(t) is a decreasing function and g(0) = 1.

[0125] In the component analysis method provided by the embodiments of the present application, since the distance field image can be used to characterize the distance of the pixel point from the lumen contour and the wall contour, and thus the prior of each position can be increased through the distance field image to assist in the segmentation of the components. Therefore, by combining the distance field image for component analysis, the accuracy and efficiency of component analysis can be further improved.

[0126] In one embodiment, the component analysis method may further include:

[0127] Performing statistical analysis on the component distribution images at N levels to obtain the component analysis result of the target vascular plaque.

[0128] In the embodiments of the present application, after obtaining the component distribution images of the target vascular plaque at N levels, statistical analysis can be performed based on these N component distribution images to obtain the component analysis result of the target vascular plaque. Among them, the statistical analysis may include analysis operations such as the maximum area analysis of each component, the volume analysis of each component, and the volume ratio analysis of each component. Correspondingly, the component analysis result may include the maximum area of each component, the volume of each component, the volume ratio of each component, etc.

[0129] Exemplarily, for any component, the number of pixel points occupied by the component in the component distribution images at each level can be counted, and then the area of the component at each level can be obtained, and further the maximum area of the component in the target vascular plaque can be determined from them. Further, the areas of the component at each level can be summed to obtain the volume of the component in the target vascular plaque. Still further, the volume ratio of the component can be calculated by taking the ratio of the volume of the component in the target vascular plaque to the total volume of the target vascular plaque.

[0130] The component analysis method provided in the embodiment of the present application can automatically generate a component distribution image in the target vascular plaque, perform statistical analysis processing, and obtain statistical analysis results of each component in the target vascular plaque. This can avoid the influence of subjective factors when the human eye performs comparison and analysis, and can improve the accuracy and efficiency of statistical analysis.

[0131] In one embodiment, the present application embodiment can implement the above step 106 by a pre-trained component analysis network to obtain a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image and the target vascular cross-sectional image of each target image at the i-th level. The component analysis method can also include:

[0132] According to a preset first training set, the analysis network is trained to obtain a component analysis network;

[0133] The first training set includes multiple first sample groups, which include multiple sequence sample images, N layers of sample blood vessel cross-sectional images corresponding to the multiple sequence sample images, and annotated component distribution images of each layer of sample blood vessel cross-sectional images.

[0134] In the embodiment of the present application, in the process of training the component analysis network, a first training set may be preset, and the first training set may include multiple first sample groups, and the first sample groups may include multiple sequence sample images, vascular cross-sectional images corresponding to the multiple sequence sample images at N levels, and annotated component distribution images of each layer of sample vascular cross-sectional images. Exemplarily, each component in the target vascular plaque in each layer of sample vascular cross-sectional images may be determined in advance based on prior knowledge, and each component may be annotated to obtain an annotated component distribution image.

[0135] In one example, input information of a component analysis network can be constructed based on multiple sequence sample images and / or vascular cross-sectional images corresponding to the multiple sequence sample images at N levels. The input information may include a sample component probability image corresponding to the i-th level, and at least one of a sample vascular cross-sectional image corresponding to the multiple sequence sample images at the i-th level and a sample histogram equalization image corresponding to each sample vascular cross-sectional image, or may also include at least one of a plurality of sample intensity contrast images and a sample distance field image corresponding to the multiple sequence sample images at the i-th level.

[0136] Among them, the determination process of the above-mentioned sample component probability image, histogram equalization image, sample intensity contrast image and sample distance field image can refer to the relevant description of the above-mentioned embodiment, and the embodiment of the present disclosure will not be repeated here. In an example, the above-mentioned sample component probability image, histogram equalization image, sample intensity contrast image and sample distance field image can be obtained by image processing based on the sample group, or can be obtained in advance through image processing and then included in the sample group, and the embodiment of the present application does not make specific limitations on this.

[0137] In this example, the above input information can be input into the analysis network for component analysis to obtain a predicted component distribution image, wherein the analysis network can be a deep segmentation network. The prediction loss of the analysis network is determined based on the predicted component distribution image and the labeled component distribution image. When the prediction loss does not meet the training requirements (for example, the prediction loss is greater than a preset loss threshold), the network parameters of the analysis network can be adjusted until the prediction loss of the analysis network meets the training requirements (for example, the prediction loss is less than or equal to the preset loss threshold), and the network training is stopped to obtain a trained component analysis network.

[0138] The component analysis method provided in the embodiment of the present application can perform component analysis processing through a component analysis network to obtain a component distribution image of the target vascular plaque, which can avoid the influence of subjectivity when performing component analysis through the human eye. The component analysis result is more objective and accurate, which can improve the accuracy and efficiency of the component analysis.

[0139] In one embodiment, in the embodiment of the present application, the above step 402 can be implemented by a segmentation network to perform segmentation processing according to the multiple sequence images to obtain the region of interest ROI image corresponding to the multiple sequence images, and the method further includes:

[0140] According to the preset second training set, the initial segmentation network is trained to obtain a segmentation network;

[0141] The second training set includes a plurality of second sample groups, and the second sample groups include multiple sequence sample images and annotated ROI images corresponding to the multiple sequence sample images.

[0142] In an embodiment of the present application, a second training set may be preset, and the second training set may include a multi-sequence sample image and a corresponding annotated ROI image in the multi-sequence sample image, wherein the annotated ROI image may be an image in which the ROI is annotated in advance in at least one sample image in the multi-sequence sample image based on prior knowledge. At least one sample image in the multi-sequence sample image may be input into the initial segmentation network to obtain a predicted ROI image, and then the prediction loss of the initial segmentation network is determined based on the predicted ROI image and the annotated ROI image. When the prediction loss does not meet the training requirements (for example, the prediction loss is greater than a preset loss threshold), the network parameters of the initial segmentation network may be adjusted until the prediction loss of the initial segmentation network meets the training requirements (for example, the prediction loss is less than or equal to a preset loss threshold), and the network training is stopped to obtain a trained or segmentation network.

[0143] The component analysis method provided in the embodiment of the present application can automatically extract the ROI area through the segmentation network to perform signal high and low comparison, and perform component analysis processing to obtain a component distribution image of the target vascular plaque. This can avoid the influence of subjectivity when performing component analysis through the human eye, and the component analysis results are more objective and accurate, which can improve the accuracy and efficiency of component analysis.

[0144] In order to enable those skilled in the art to better understand the embodiments of the present application, the embodiments of the present application are described below through specific examples.

[0145] After acquiring multiple sequence images, image registration is performed to achieve spatial alignment of each image in the multiple sequence images (the images mentioned below are all registered images).

[0146] Blood vessel centerline extraction based on multiple sequence images, including but not limited to centerline extraction methods based on single images and multiple sequence images, based on traditional algorithms and based on deep learning, and manual extraction methods. Blood vessel plaque detection based on multiple sequence images, including but not limited to plaque extraction methods based on single images and multiple sequence images, based on traditional algorithms and based on deep learning, and manual extraction methods that manually determine the starting and ending points of plaques on the centerline.

[0147] Based on the above vascular plaque detection results, the following operations are performed within the range of the target vascular plaque:

[0148] At the blood vessel centerline point P0 of the image Si in the multi-series image, k points are acquired along the front-to-back direction of the blood vessel on the blood vessel centerline with a step size of λ, and the (2k+1) points acquired are used as the center to reconstruct (2k+1) blood vessel cross-sectional images Ii.

[0149] A multi-channel three-dimensional input information is constructed based on at least one image in the above-mentioned multiple sequence images, and the input information is input into a pre-trained segmentation model to obtain a ROI image.

[0150] The ROI image and image Si are used to calculate the statistical image Vi of each ROI area in the image Si (the statistical image can be obtained by calculating the mean or median), and each pixel point in the vascular cross-sectional image Ii is divided by each pixel point in Vi to obtain the intensity relative image Di corresponding to the image Si.

[0151] For the above-mentioned blood vessel cross-sectional image Ii, a depth segmentation model or manual editing by the user is used to obtain a lumen and wall segmentation image W1 corresponding to the blood vessel cross-sectional image. In the lumen and wall segmentation image W1, the background, lumen, and wall can be represented by different values ​​respectively.

[0152] The lumen contour L1 and the tube wall contour L2 are obtained from the lumen-tube wall segmentation image W1. For each pixel point inside the tube wall in the lumen-tube wall segmentation image W1, the distance x1 to the lumen contour L1 and the distance x2 to the tube wall contour L2 are calculated. The distance field result f(x1, x2) of the current point is obtained using the given distance field mapping function f. The distance field image F is constructed according to the distance field results corresponding to each pixel point.

[0153] Using the component prior distribution table T, the relative distribution sequence of each component on different images in the multi-sequence image is obtained. According to the given probability prior function P, the component prior distribution table T and each intensity relative image Di are probability-converted to obtain the probability of each pixel being each component. The probability prior function P is called for all pixel points on the tube wall to calculate the probability, and the probability of the pixel point outside the tube wall being a certain component is 0. The component probability map Pi of the component is constructed based on the probability of all pixel points being a certain component.

[0154] Histogram equalization is performed on the above-mentioned blood vessel cross-sectional images Ii to obtain a histogram equalized image Hi corresponding to each blood vessel cross-sectional image Ii.

[0155] The above Ii, Di, Hi, F, Pi are used as multi-channel or multi-branch inputs and the pre-trained component analysis network (for example, the network structure of the component analysis network can refer to Figure 6a or Figure 6b ), and obtain a component distribution image. Based on the component distribution image, the components of the target vascular plaque are counted to obtain statistical information of each component in the target vascular plaque, such as the maximum area, volume, volume proportion, etc.

[0156] Based on the component analysis method provided in the embodiment of the present application, the reference area ROI can be automatically obtained, and the intensity relative diagram of the signal value of the vascular wall area of ​​the multi-sequence image and the reference area can be obtained. The signal difference between the vascular wall area and the reference area can be quantitatively compared, which can avoid the errors caused by the doctor's eye positioning and subjective judgment. Based on the original image and the prior features such as the intensity relative image, the distance field image, and the component probability image, the deep learning method is used to automatically segment the plaque components and obtain the component analysis results of each component of the plaque. Based on the analysis results of each component, statistical information such as the volume, volume proportion, and distribution position of each component of the plaque can be obtained, which can improve the accuracy and efficiency of component analysis.

[0157] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0158] Based on the same inventive concept, the embodiment of the present application also provides a component analysis device for implementing the component analysis method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more component analysis device embodiments provided below can refer to the limitations on the component analysis method above, and will not be repeated here.

[0159] In one embodiment, Figure 7 As shown, a component analysis device is provided, including: a creation module 702, a first determination module 704 and a first processing module 706, wherein:

[0160] A creation module 702 is used to create N vascular cross-sectional images on N levels for any target image in the multiple sequence images according to the vascular centerline and target vascular plaque of the target image, where N is an integer greater than 0;

[0161] A first determination module 704 is used to determine a component probability image of the target vascular plaque at the i-th level according to the target vascular cross-sectional image at the i-th level of each of the target images, wherein i is an integer greater than 0 and less than or equal to N;

[0162] The first processing module 706 is used to obtain a component distribution image corresponding to the target blood vessel plaque at the i-th level according to the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level.

[0163] A component analysis device provided in an embodiment of the present application creates N cross-sectional images of blood vessels on N levels according to the blood vessel centerline and target blood vessel plaque of the target image for any target image in a multi-sequence image, wherein N is an integer greater than 0; determines the component probability image of the target blood vessel plaque on the i-th level according to the cross-sectional images of the target blood vessels of each target image on the i-th level; obtains the component distribution image corresponding to the target blood vessel plaque on the i-th level according to the component probability image and the cross-sectional images of the target blood vessels of each target image on the i-th level. Based on the component analysis device provided in an embodiment of the present application, the component probability image of the target blood vessel plaque can be obtained by analysis, and then the component probability image can be automatically and objectively analyzed to obtain the component distribution image of the target blood vessel plaque, which can avoid the influence of subjectivity when performing component analysis by human eyes, making the component analysis result more objective and accurate, and improving the accuracy and efficiency of component analysis.

[0164] In one embodiment, the first processing module 706 is further configured to:

[0165] Obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image and the vascular cross-sectional image of each target image at the i-th level;

[0166] The blood vessel cross-sectional image includes at least one of the target blood vessel cross-sectional image and a histogram-equalized image obtained by performing histogram-equalization processing on the target blood vessel cross-sectional image.

[0167] In one embodiment, the first processing module 706 is further configured to:

[0168] Obtaining a correlation analysis image corresponding to the target blood vessel plaque according to each of the target images and / or the target blood vessel cross-sectional image of each of the target images at the i-th level;

[0169] Obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image, the vascular cross-sectional image of each target image at the i-th level, and the correlation analysis image;

[0170] The correlation analysis image includes at least one of a distance field image corresponding to the target blood vessel cross-sectional image and an intensity contrast image corresponding to each of the target blood vessel cross-sectional images.

[0171] In one embodiment, the first determining module 702 is further configured to:

[0172] Perform segmentation processing according to the multiple sequence images to obtain a region of interest ROI image corresponding to the multiple sequence images;

[0173] Obtaining statistical images corresponding to each of the target images according to the ROI image and the corresponding area of ​​the ROI image in each of the target images;

[0174] For any of the target images, an intensity contrast image is obtained according to the target blood vessel cross-sectional image on the i-th layer and the statistical image corresponding to the target image, wherein the intensity contrast image is used to represent the signal intensity contrast result between the target blood vessel cross-sectional image and the statistical image;

[0175] According to the intensity contrast image and component distribution prior information corresponding to each of the target images, a component probability image of the target vascular plaque at the i-th level is obtained.

[0176] In one embodiment, the device further comprises:

[0177] A second processing module is used to obtain a lumen wall segmentation image of the multiple sequence images at the i-th level according to a cross-sectional image of a target blood vessel at the i-th level in at least one of the target images;

[0178] A second determination module is used to determine a first distance between each pixel point inside the tube wall and the tube lumen contour in the tube lumen and tube wall segmentation image, and a second distance between each pixel point and the tube wall contour;

[0179] The third processing module is used to obtain the distance field image on the i-th layer according to the first distance and the second distance corresponding to each pixel point, and the pixel value of any pixel point in the distance field image is used to represent the distance between the pixel point and the lumen contour and the tube wall contour.

[0180] In one embodiment, the device further comprises:

[0181] The fourth processing module is used to perform statistical analysis based on the component distribution images at N levels to obtain component analysis results of the target vascular plaque.

[0182] In one embodiment, the device further comprises:

[0183] A first training module, used for training the analysis network according to a preset first training set to obtain the component analysis network;

[0184] The first training set includes multiple first sample groups, which include multiple sequence sample images, N layers of sample blood vessel cross-sectional images corresponding to the multiple sequence sample images, and annotated component distribution images of each layer of the sample blood vessel cross-sectional images.

[0185] In one embodiment, the device further comprises:

[0186] A second training module, used for training the initial segmentation network according to a preset second training set to obtain the segmentation network;

[0187] The second training set includes a plurality of second sample groups, and the second sample groups include multiple sequence sample images and annotated ROI images corresponding to the multiple sequence sample images.

[0188] Each module in the above-mentioned component analysis device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0189] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a component analysis method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0190] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0191] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0192] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0193] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0194] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0195] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0196] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0197] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A component analysis method, characterized in that: The method comprises: For any target image in the multiple sequence images, create N vascular cross-sectional images on N levels according to the vascular centerline and target vascular plaque of the target image; According to the target blood vessel cross-sectional image on the i-th layer and the statistical image corresponding to each of the target images, an intensity contrast image corresponding to each of the target images is obtained; the intensity contrast image is used to characterize the signal intensity contrast result between the target blood vessel cross-sectional image and the statistical image; Performing a three-value conversion process on the intensity contrast images corresponding to each of the target images by using a threshold value to obtain a three-value sequence corresponding to each pixel point of the intensity contrast image; According to the signal intensity level corresponding to each component on each of the target images, a relative distribution sequence corresponding to each component is obtained; Calculating the distance between the three-value sequence corresponding to each pixel of the intensity contrast image and the relative distribution sequence corresponding to each component, and converting the distance corresponding to each pixel into the probability that each pixel is each component; According to the probability that all pixels are each component, a component probability image of the target vascular plaque at the i-th layer is obtained; A component distribution image corresponding to the target blood vessel plaque at the i-th level is obtained according to the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level.

2. The method according to claim 1, characterized in that The step of obtaining a component distribution image corresponding to the target blood vessel plaque at the i-th level according to the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level includes: Obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image and the vascular cross-sectional image of each target image at the i-th level; The blood vessel cross-sectional image includes at least one of the target blood vessel cross-sectional image and a histogram-equalized image obtained by performing histogram-equalization processing on the target blood vessel cross-sectional image.

3. The method according to claim 2, characterized in that The step of obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image and the vascular cross-sectional image of each target image at the i-th level includes: Obtaining a correlation analysis image corresponding to the target blood vessel plaque according to each of the target images and / or the target blood vessel cross-sectional image of each of the target images at the i-th level; Obtaining a component distribution image corresponding to the target vascular plaque at the i-th level according to the component probability image, the vascular cross-sectional image of each target image at the i-th level, and the correlation analysis image; The correlation analysis image includes at least one of a distance field image corresponding to the target blood vessel cross-sectional image and an intensity contrast image corresponding to each of the target blood vessel cross-sectional images.

4. The method according to claim 1, characterized in that: Obtaining statistical images corresponding to each of the target images, including: Perform segmentation processing according to the multiple sequence images to obtain a region of interest ROI image corresponding to the multiple sequence images; According to the ROI image and the corresponding area of ​​the ROI image in each of the target images, a statistical image corresponding to each of the target images is obtained.

5. The method according to claim 3, characterized in that: The method further comprises: According to a cross-sectional image of a target blood vessel at an i-th level in at least one of the target images, a lumen and wall segmentation image of the multiple sequence images at the i-th level is obtained; Determine a first distance between each pixel point inside the tube wall and the tube lumen contour in the tube lumen and tube wall segmentation image, and a second distance between each pixel point and the tube wall contour; According to the first distance and the second distance corresponding to each pixel point, a distance field image on the i-th layer is obtained, and the pixel value of any pixel point in the distance field image is used to characterize the distance between the pixel point and the lumen contour and the tube wall contour.

6. The method according to any one of claims 1 to 5, characterized in that The method further comprises: Statistical analysis is performed on the component distribution images at N levels to obtain component analysis results of the target vascular plaque.

7. The method according to any one of claims 1 to 5, characterized in that The method uses a component analysis network to obtain a component distribution image corresponding to the target vascular plaque at the i-th level based on the component probability image and the target blood vessel cross-sectional image at the i-th level of each target image, and the method further includes: According to a preset first training set, the analysis network is trained to obtain the component analysis network; The first training set includes multiple first sample groups, which include multiple sequence sample images, N layers of sample blood vessel cross-sectional images corresponding to the multiple sequence sample images, and annotated component distribution images of each layer of the sample blood vessel cross-sectional images.

8. The method according to claim 4, characterized in that The method implements the segmentation processing according to the multiple sequence images through a segmentation network to obtain a region of interest ROI image corresponding to the multiple sequence images, and the method further includes: Training the initial segmentation network according to a preset second training set to obtain the segmentation network; The second training set includes a plurality of second sample groups, and the second sample groups include multiple sequence sample images and annotated ROI images corresponding to the multiple sequence sample images.

9. A component analysis device, characterized in that: The device comprises: A creation module, for creating N vascular cross-sectional images on N levels according to the vascular centerline and target vascular plaque of any target image in the multiple sequence images, wherein N is an integer greater than 0; The first determination module is used to obtain an intensity contrast image corresponding to each target image based on the target blood vessel cross-sectional image on the i-th level and the statistical image corresponding to each target image; the intensity contrast image is used to characterize the signal intensity contrast result between the target blood vessel cross-sectional image and the statistical image; the intensity contrast image corresponding to each target image is subjected to ternary conversion processing by a threshold value to obtain a three-value sequence corresponding to each pixel of the intensity contrast image; according to the signal intensity level corresponding to each component on each target image, a relative distribution sequence corresponding to each component is obtained; the distance from the three-value sequence corresponding to each pixel of the intensity contrast image to the relative distribution sequence corresponding to each component is calculated, and the distance corresponding to each pixel is converted into the probability of each pixel being each component; according to the probability of all pixels being each component, a component probability image of the target blood vessel plaque on the i-th level is obtained; wherein i is an integer greater than 0 and less than or equal to N; The first processing module is used to obtain a component distribution image corresponding to the target blood vessel plaque at the i-th level according to the component probability image and the target blood vessel cross-sectional image of each target image at the i-th level.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images