Image reconstruction method and device, computer equipment and storage medium

Abstract

The invention relates to an image reconstruction method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring scanning data of all phases and reconstructing to obtain images corresponding to a plurality of phases as to-be-evaluated images; calculating a quality index of each to-be-evaluated image according to an image quality evaluation rule; and according to the quality index of each to-be-evaluated image, calculating to obtain an optimal phase, and obtaining an optimal phase image. The method does not depend on interaction of an artificial interface, can automatically detect and extract interested blood vessels, and automatically analyzes blood vessel image quality. The coronary artery reconstruction process is simplified, and the time for a doctor to evaluate image selection parameters is saved.

Description

[0001] This application is a divisional application filed with the State Intellectual Property Office on September 27, 2018, with application number 2018111343751, entitled "Image Reconstruction Method, Apparatus, Computer Equipment and Storage Medium", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of medical device technology, and in particular to an image reconstruction method, apparatus, computer device, and storage medium. Background Technology

[0003] Angiography is an auxiliary diagnostic technique. In this era of advanced technology, angiography is widely used in the diagnosis and treatment of various diseases in clinical practice. It helps doctors detect diseases in a timely manner, control disease progression, and effectively improve patient survival rates. Therefore, the image quality of angiography is crucial for diagnosis.

[0004] Current traditional techniques for evaluating the quality of coronary angiography in clinical practice involve manual observation and subjective assessment of reconstructed images. During coronary angiography, the heart's pulsation produces motion artifacts, requiring physicians to select an appropriate reconstruction phase to obtain a diagnostic image. In traditional manual assessment, the computer generates a series of images for evaluation through a user interface, which the physician then manually reviews and selects a specific reconstruction phase for reconstructing. This assessment method not only complicates the reconstruction process but also increases the burden on physicians in evaluating and selecting image quality, and may lead to repeated reconstruction and evaluation. Summary of the Invention

[0005] Therefore, it is necessary to provide an image quality evaluation method and image reconstruction method, apparatus, computer equipment, and storage medium that can automatically perform optimal phase selection to address the above-mentioned technical problems.

[0006] An image quality assessment method includes: acquiring multiple images to be evaluated; obtaining a blood vessel image of interest based on the images to be evaluated and a segmentation threshold; and performing image quality assessment based on the blood vessel image of interest.

[0007] In one embodiment, the image quality evaluation based on the image of the vessel of interest includes: calculating the quality index of the corresponding image to be evaluated based on the image of the vessel of interest; and performing image quality evaluation based on the quality index of the image to be evaluated.

[0008] In one embodiment, calculating the quality index of the corresponding image to be evaluated based on the image of the vessel of interest includes: calculating the morphological regularity of the corresponding image to be evaluated based on the image of the vessel of interest; calculating the edge sharpness of the corresponding image to be evaluated based on the boundary of the image of the vessel of interest and the gradient map of the image of the vessel of interest; and calculating the quality index of each image to be evaluated based on the morphological regularity and the edge sharpness of the image to be evaluated.

[0009] In one embodiment, calculating the morphological regularity of the corresponding image to be evaluated based on the image of the vessel of interest includes: calculating the perimeter and area of ​​the target object in each image to be evaluated based on the image of the vessel of interest; and calculating the morphological regularity of the corresponding image to be evaluated based on the perimeter and area of ​​the target object in each image to be evaluated.

[0010] In one embodiment, the image quality evaluation based on the quality index of the image to be evaluated includes: selecting the image to be evaluated with the highest quality index as the image with the best image quality.

[0011] In one embodiment, the segmentation threshold is determined by selecting a preset multiple of the maximum grayscale value in the image to be evaluated as the segmentation threshold.

[0012] An image reconstruction method includes: acquiring scan data of all phases to reconstruct multiple phase-corresponding images as images to be evaluated; calculating the quality index of each image to be evaluated according to image quality evaluation rules; calculating the optimal phase based on the quality index of each image to be evaluated, and obtaining the optimal phase image.

[0013] In one embodiment, the step of acquiring scan data of all phases and reconstructing multiple phase-corresponding images as images to be evaluated includes: calculating the average optimal phase based on the multiple phase-corresponding images; selecting phase images within a preset range near the average optimal phase and extracting regions of interest (ROI) images from the selected multiple phase images; extracting the vessel centerline of the corresponding ROI image based on the multiple ROI images; and performing image segmentation within a preset range with the vessel centerline as the center to obtain multiple images to be evaluated.

[0014] In one embodiment, calculating the average optimal phase based on images corresponding to multiple phases includes: calculating cardiac motion parameters for multiple phases based on images corresponding to multiple phases; and calculating the average optimal phase based on cardiac motion parameters for multiple phases.

[0015] In one embodiment, calculating the cardiac motion parameters of all phases based on the images corresponding to multiple phases includes: calculating the average absolute difference between two adjacent phase images based on the pixel values ​​of the images corresponding to two adjacent phases and the image matrix size; and calculating the motion parameters of multiple phases based on the average absolute difference between the pixel values ​​of two adjacent phase images.

[0016] In one embodiment, before calculating the average absolute difference between two adjacent phase images based on the pixel values ​​of the images corresponding to the adjacent two phases and the image matrix size, the method further includes: preprocessing the images corresponding to the plurality of phases, wherein the preprocessing includes: performing image segmentation on the images based on an image threshold to eliminate regions unrelated to heart movement and obtain images of regions related to heart movement.

[0017] In one embodiment, calculating the motion parameters of multiple phases based on the average absolute difference of pixel values ​​of two adjacent phase images includes: obtaining the average absolute difference between a phase image and the previous phase image, and using it as a first parameter; obtaining the average absolute difference between a phase image and the next phase image, and using it as a second parameter; and adding the first parameter and the second parameter of the same phase image to obtain the motion parameters of the corresponding phase.

[0018] In one embodiment, calculating the average optimal phase based on cardiac motion parameters of multiple phases includes: during cardiac systole, taking the phase with the smallest motion parameter among the cardiac systole phases as the average optimal phase of cardiac systole; and during cardiac diastole, taking the phase with the smallest motion parameter among the cardiac diastole phases as the average optimal phase of cardiac diastole.

[0019] In one embodiment, selecting phase images within a preset range near the average optimal phase and extracting regions of interest (ROI) images from the selected multiple phase images includes: selecting phase images within a preset range near the average optimal phase; smoothing the phase images within the preset range using a Gaussian low-pass filter; extracting ventricular region images from the smoothed phase images; calculating a contrast agent threshold based on the ventricular region images; performing image segmentation based on the ventricular region images and the contrast agent threshold to obtain a contrast agent region image; and selecting a region of interest image from the contrast agent region image.

[0020] In one embodiment, extracting the ventricular region image from the smoothed phase image includes: performing image segmentation based on the smoothed phase image and a bone threshold to obtain a bone region image; performing maximum density projection on the bone region image along the thoracic cavity axis to obtain a maximum density projection image of the bone region image; calculating the thoracic cavity contour boundary based on the maximum density projection image of the bone region image; and obtaining the ventricular region image based on the smoothed phase image and the thoracic cavity contour boundary.

[0021] In one embodiment, obtaining the ventricular region image based on the smoothed phase image and the chest cavity contour boundary includes: obtaining a chest cavity image based on the smoothed phase image and the chest cavity contour boundary; calculating connected components based on the chest cavity image, and selecting the image in the connected component with the largest number of pixels as the ventricular region image.

[0022] In one embodiment, calculating the contrast agent threshold based on the ventricular region image includes: calculating the gradient image of the ventricular region image based on the ventricular region image; using ventricular region images where the gradient image grayscale value at a corresponding position in the ventricular region image is greater than a proportional threshold as marker images; and calculating the contrast agent threshold using the Otsu algorithm based on the grayscale value of each pixel in the marker image.

[0023] In one embodiment, extracting the vessel centerline of a corresponding region of interest image based on multiple region of interest images includes: acquiring coronal and sagittal view images of the multiple region of interest images; determining the main vessel trunk based on the coronal and sagittal view images; filtering out false positive vessels based on the main vessel trunk; determining the vessel center position of each slice based on the vessel trunk after filtering out false positive vessels; and obtaining the vessel centerline of the corresponding region of interest image based on the vessel center position of each slice.

[0024] In one embodiment, filtering false positive vessels based on the main vascular trunk includes: filtering non-main trunk vessels based on the main vascular trunk; and filtering the main trunk vessels based on the main vascular trunk after filtering non-main trunk vessels.

[0025] In one embodiment, the step of segmenting the image within a preset range centered on the vessel centerline to obtain multiple images to be evaluated includes: performing a top-hat transform on the region of interest image to obtain a region of interest image highlighting the target object; segmenting the region of interest image that preserves the intraventricular region based on a soft tissue threshold; and selecting the region of interest image that preserves the intraventricular region within a preset range centered on the vessel centerline as the image to be evaluated for the corresponding phase.

[0026] In one embodiment, calculating the quality index of each image to be evaluated according to the image quality evaluation rules includes: obtaining the image of interest (ROI) based on the image to be evaluated and the segmentation threshold; and calculating the quality index of the corresponding image to be evaluated based on the ROI image.

[0027] In one embodiment, the segmentation threshold is determined by selecting a preset multiple of the maximum grayscale value in the image to be evaluated as the segmentation threshold.

[0028] In one embodiment, calculating the quality index of the corresponding image to be evaluated based on the image of the vessel of interest includes: calculating the perimeter and area of ​​the target object in each image to be evaluated based on the image of the vessel of interest; calculating the morphological regularity of the corresponding image to be evaluated based on the perimeter and area of ​​the target object in each image to be evaluated; calculating the edge sharpness of the corresponding image to be evaluated based on the boundary of the image of the vessel of interest and the gradient map of the image of the vessel of interest; and calculating the quality index of each image to be evaluated based on the morphological regularity and the edge sharpness of the image to be evaluated.

[0029] An image quality assessment device includes: an acquisition module for acquiring multiple images to be assessed; a vessel of interest image extraction module for obtaining a vessel of interest image based on the images to be assessed and a segmentation threshold; and an image quality assessment module for performing image quality assessment based on the vessel of interest image.

[0030] An image reconstruction apparatus includes: an image selection module for acquiring scan data of all phases to reconstruct multiple phase images as images to be evaluated; a quality index calculation module for calculating the quality index of each image to be evaluated according to image quality evaluation rules; and an image reconstruction module for calculating the optimal phase and obtaining the optimal phase image based on the quality index of each image to be evaluated.

[0031] A computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of any of the methods described above.

[0032] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0033] The aforementioned image quality assessment and reconstruction methods, apparatus, computer equipment, and storage medium first use a preset multiple of the maximum grayscale value of the pixels in the image to be evaluated as a segmentation threshold. Then, the image to be evaluated is segmented using this threshold to obtain images of the vessels of interest. Based on these images, a quality index is calculated for the corresponding images to be evaluated. Image quality is then assessed for all images based on this quality index. The system automatically evaluates images of corresponding phases according to image quality assessment rules, simplifying the reconstruction process, reducing the burden on doctors in image evaluation, avoiding redundant image evaluations, further improving the accuracy of optimal phase selection, and ultimately enhancing reconstruction quality. Attached Figure Description

[0034] Figure 1 This is a flowchart illustrating an image quality evaluation method in one embodiment;

[0035] Figure 2 This is a flowchart illustrating an image reconstruction method in one embodiment;

[0036] Figure 3 This is a flowchart illustrating the image reconstruction method in another embodiment;

[0037] Figure 4 This is a flowchart illustrating a method for calculating the average optimal phase in one embodiment;

[0038] Figure 5 This is a flowchart illustrating a method for extracting a region of interest image in one embodiment;

[0039] Figure 6 This is a flowchart illustrating a method for extracting images of the ventricular region in one embodiment;

[0040] Figure 7 This is a flowchart illustrating a method for extracting the center line of blood vessels from an image of a region of interest in one embodiment.

[0041] Figure 8 This is a flowchart illustrating a method for extracting an image to be evaluated in one embodiment;

[0042] Figure 9 This is a flowchart illustrating a method for calculating the quality index of an image to be evaluated in one embodiment.

[0043] Figure 10 This is a structural block diagram of an image reconstruction apparatus in one embodiment;

[0044] Figure 11 This is an internal structural diagram of a computer device in one embodiment.

[0045] Figure description: Image selection module 100, quality index calculation module 200, image reconstruction module 300. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0047] Computed tomography (CT) equipment typically includes a gantry, a scanning bed, and a control console for the physician. An X-ray tube is located on one side of the gantry, and a detector is located on the opposite side. The control console is a computer device that controls the scan, receives the scan data acquired by the detector, processes and reconstructs the data, and ultimately forms a CT image. During a CT scan, the patient lies on the scanning bed, which guides the patient into the aperture of the gantry. The X-ray tube on the gantry emits X-rays, which pass through the patient and are received by the detector to form scan data. This scan data is then transmitted to the computer device, which performs preliminary processing and image reconstruction to obtain the CT image.

[0048] In one embodiment, such as Figure 1 As shown, an image quality assessment method is provided, including the following steps:

[0049] Step S102: Obtain multiple images to be evaluated.

[0050] Specifically, multiple images to be evaluated are acquired, and a preset multiple of the maximum grayscale value in each image is selected as the segmentation threshold. During normal CT scanning, the object being scanned is continuously scanned over a period of time, yielding corresponding scan data. Multiple images to be evaluated are acquired based on this scan data. For each image, a preset multiple of the maximum value of its individual pixels is used as the segmentation threshold. Before obtaining the segmentation threshold, the resolution of the images to be evaluated can be increased. Increasing the resolution improves the accuracy of calculating blood vessel morphology and edges; two-dimensional image interpolation can be used to enhance the resolution. At least one preset multiple is used.

[0051] Step S104: Based on the image to be evaluated and the segmentation threshold, obtain the image of the blood vessel of interest.

[0052] Specifically, images whose grayscale values ​​are greater than a segmentation threshold are designated as vessels of interest (ROIs) for the corresponding segmentation threshold. Multiple segmentation thresholds result in multiple ROIs. Preferably, three segmentation thresholds are obtained based on three preset multiples of the maximum grayscale value. The first segmentation threshold is used to segment the image, identifying regions where pixel grayscale values ​​are greater than the first threshold as the first ROI image; the second segmentation threshold is used, identifying regions where pixel grayscale values ​​are greater than the second threshold as the second ROI image; and the third segmentation threshold is used, identifying regions where pixel grayscale values ​​are greater than the third threshold as the third ROI image.

[0053] Step S106: Evaluate the image quality based on the image of the vessel of interest.

[0054] Specifically, based on the image of the blood vessel of interest, the quality index of the corresponding image to be evaluated is calculated; and based on the quality index of the image to be evaluated, the image quality is evaluated.

[0055] More specifically, based on the image of the vessel of interest, the quality index of the corresponding image to be evaluated is calculated. First, the morphological regularity of the corresponding image to be evaluated is calculated based on the image of the vessel of interest. More specifically, based on the image of the vessel of interest, the perimeter and area of ​​the target object in each image to be evaluated are calculated, that is, the perimeter and area of ​​the vessel in each image of the vessel of interest are calculated separately. Then, based on the perimeter and area of ​​the target object in each image to be evaluated, the morphological regularity of the corresponding image to be evaluated is calculated. Then, based on the boundary and gradient map of the image of the vessel of interest, the edge sharpness of the corresponding image to be evaluated is calculated. Finally, based on the morphological regularity and edge sharpness of the images to be evaluated, the quality index of each image to be evaluated is calculated. The number of vessels at the same location in different phases may be inconsistent. Considering that subsequent comparisons need to be made on the same benchmark, that is, the number of vessels in the images at the same location in each phase should be consistent, a reference quantity, the base number of vessels, is introduced here. Based on the base number of vessels and the number of vessels in the images to be evaluated, the morphological regularity matrix and the edge sharpness matrix of the images to be evaluated in each phase are obtained. Since the magnitudes of regularity of shape and edge sharpness are different, it is necessary to bring the two measures to the same baseline. This can be done by weighting or by normalizing the regularity of shape and edge sharpness.

[0056] More specifically, image quality is evaluated based on the quality index of the image to be evaluated, and the image with the highest quality index is selected as the image with the best image quality.

[0057] The aforementioned image quality assessment method first uses a preset multiple of the maximum grayscale value of the pixels in the image to be evaluated as a segmentation threshold. Then, the image to be evaluated is segmented using this threshold to obtain images of the vessels of interest. Based on these images, a quality index is calculated for the corresponding images to be evaluated. The image quality is then assessed for all images based on this quality index. This automated evaluation of image quality based on the quality index reduces the burden on doctors in selecting images and further minimizes the need for repeated reconstruction of a single image.

[0058] In one embodiment, such as Figure 2 As shown, an image reconstruction method is provided, including the following steps:

[0059] Step S202: Obtain the scan data of all phases and reconstruct multiple phase images as images to be evaluated.

[0060] Specifically, during a normal CT scan, the object being scanned is continuously scanned over a period of time, and corresponding scan data is obtained. Within a cardiac cycle, each phase corresponds to data acquired by the CT scan; that is, data is available for 100 phases ranging from 1% to 100% within each cardiac cycle. Phase images for each phase are reconstructed based on the data from all phases. Reconstruction can use small-matrix thick-slice reconstruction, but this has low resolution and can negatively impact subsequent segmentation. Alternatively, precise reconstruction can be used, automatically locating the region of interest (ROI) as the reconstruction center and employing small-field-of-view, small-matrix thick-slice reconstruction. First, the average optimal phase is calculated based on the images corresponding to all phases. Phase images within a preset range near the average optimal phase are selected, and the ROI image is extracted from the selected multiple phase images. The vessel centerline is then extracted from the ROI images. Finally, image segmentation is performed within a preset range centered on the vessel centerline, resulting in multiple images to be evaluated.

[0061] Step S204: Calculate the quality index of each image to be evaluated according to the image quality evaluation rules.

[0062] Specifically, a preset multiple of the maximum grayscale value is first selected as the segmentation threshold among multiple images to be evaluated. The images to be evaluated are then segmented according to the segmentation threshold to obtain multiple images of vessels of interest. Based on the multiple images of interest for each image to be evaluated, the quality index of the corresponding image is calculated. This yields the quality index of the images to be evaluated for all phases within each cardiac cycle.

[0063] Step S206: Calculate the optimal phase based on the quality index of each image to be evaluated, and obtain the optimal phase image.

[0064] Specifically, within each cardiac cycle, the image to be evaluated with the highest quality index is selected, and the phase of the image with the highest quality index is taken as the optimal phase. The optimal phase is then reconstructed to obtain the optimal phase image. Alternatively, the scan data of all phases can be reconstructed to obtain images of all phases. The image to be evaluated with the highest quality index is selected, and the phase of the image with the highest quality index is taken as the optimal phase. Then, the corresponding phase image is selected as the optimal phase image.

[0065] The image reconstruction method described above reconstructs the scan description data for all phases, obtaining images for all phases. The corresponding phase image to be evaluated is then obtained from these images. Based on image quality evaluation rules, a quality index is calculated for each image to be evaluated. The optimal phase is then determined based on the quality indices of all phase images to be evaluated, and finally, the optimal phase image is reconstructed. This method does not rely on a human interface and can automatically detect and extract the image to be evaluated, and automatically analyze the quality of the vascular image. In coronary angiography, it can be used to automatically select the optimal phase, eliminating the need for physicians to evaluate images and select the reconstruction phase, thus simplifying the coronary reconstruction process and saving physicians time in evaluating images and selecting parameters.

[0066] In one embodiment, such as Figure 3 As shown, another image reconstruction method is provided, which includes the following steps:

[0067] Step S302: Obtain scan data for all phases and reconstruct images corresponding to all phases.

[0068] Specifically, during a normal CT scan, the object being scanned is continuously scanned over a period of time, and the corresponding scan data is obtained. Within a cardiac cycle, each phase corresponds to data acquired by the CT scan; that is, data is available for 1%-100% of the 100 phases within each cardiac cycle. Based on the data from all phases, the corresponding phase images are reconstructed.

[0069] Step S304: Calculate the average optimal phase based on the images corresponding to all phases.

[0070] Specifically, firstly, cardiac motion parameters for all phases are calculated based on the images corresponding to all phases. Then, the average optimal phase is calculated based on the cardiac motion parameters for all phases. The average optimal phase can be either the average optimal phase during cardiac systole or the average optimal phase during cardiac diastole.

[0071] Step S306: Select phase images within a preset range near the average optimal phase, and extract the region of interest image from the selected multiple phase images.

[0072] Specifically, a phase image within a preset range near the average optimal phase is selected, and then a Gaussian low-pass filter is used to smooth the phase image within the preset range. The ventricular region image is extracted from the smoothed phase image, the contrast agent threshold is calculated based on the ventricular region image, and the ventricular region image is segmented according to the contrast agent threshold to obtain the contrast agent region image. The region of interest is then selected from the contrast agent region image.

[0073] Step S308: Extract the blood vessel centerline from the corresponding region of interest image based on multiple region of interest images.

[0074] Specifically, coronal and sagittal view images of multiple regions of interest are acquired. Based on the coronal and sagittal view images, the main blood vessels are determined, false positive vessels in the main blood vessels are filtered out, and the center position of the blood vessels in each slice is determined. Based on the center position of the blood vessels in each slice, the center line of the blood vessels in the corresponding region of interest image is obtained.

[0075] Step S310: Using the blood vessel centerline as the center, perform image segmentation within a preset range to obtain multiple images to be evaluated.

[0076] Specifically, a top-hat transform is performed on the region of interest image to obtain a region of interest image highlighting the target object. Based on the soft tissue threshold, a region of interest image preserving the intraventricular region is obtained. The region of interest image preserving the intraventricular region within a preset range centered on the vessel centerline is selected as the image to be evaluated for the corresponding phase.

[0077] Step S312: Calculate the quality index of each image to be evaluated according to the image quality evaluation rules.

[0078] Specifically, a preset multiple of the maximum grayscale value is first selected as the segmentation threshold among multiple images to be evaluated. The images to be evaluated are then segmented according to the segmentation threshold to obtain multiple images of vessels of interest. Based on the multiple images of interest for each image to be evaluated, the quality index of the corresponding image is calculated. This yields the quality index of the images to be evaluated for all phases within each cardiac cycle.

[0079] Step S314: Calculate the optimal phase based on the quality index of each image to be evaluated, and reconstruct the optimal phase image.

[0080] Specifically, within each cardiac cycle, the image to be evaluated with the highest quality index is selected, and the phase of the image to be evaluated with the highest quality index is taken as the optimal phase. The optimal phase is then reconstructed to obtain the optimal phase image.

[0081] The image reconstruction method described above can eliminate the interference of cardiac motion and use the region of interest as the image to be evaluated, enabling more accurate automatic evaluation and further improving the accuracy of automatic evaluation, thus saving doctors more time in selecting parameters for image evaluation.

[0082] In one embodiment, such as Figure 4 As shown, a method for calculating the average optimal phase is provided, including the following steps:

[0083] Step S402: Calculate the average absolute difference between two adjacent phase images based on the pixel values ​​of the images corresponding to the two adjacent phases and the size of the image matrix.

[0084] Specifically, in cardiac reconstruction, the phase range is from 1% to 100%, serving as a preliminary positioning tool. Evaluating all phase images would reduce efficiency; therefore, a phase selection range needs to be defined within this range for subsequent image evaluation. Furthermore, the three-dimensional images obtained from multiplanar reconstruction require continuous blood vessels. If the phase selection span is too large, the optimal phase image sequence between different cardiac cycles may become discontinuous.

[0085] The average optimal phase can be determined through clinical experience, such as 45% during cardiac systole and 75% during cardiac diastole.

[0086] The average optimal phase can also be calculated. Before calculating the average absolute difference between two adjacent phase images, the images corresponding to multiple phases are preprocessed. The preprocessing includes: performing image segmentation on the sampled phase images according to the image threshold, eliminating regions that are not related to heart motion, and obtaining images of regions related to heart motion.

[0087] In one embodiment, the equation for calculating the mean absolute difference between two adjacent sampled phase images can be:

[0088]

[0089] Where A and B are images of two adjacent phases; A(i,j) is the gray value of the pixel at coordinate (i,j) in image A; B(i,j) is the gray value of the pixel at coordinate (i,j) in image B; matrix is ​​the size of the image matrix; MAD(A,B) is the mean absolute difference between A and B.

[0090] Step S404: Calculate the motion parameters of each phase based on the average absolute difference of pixel values ​​between two adjacent phase images.

[0091] Specifically, the mean absolute difference between the sampled phase image and the previous sampled phase image is obtained and used as the first parameter. The mean absolute difference between the sampled phase image and the subsequent sampled phase image is obtained and used as the second parameter. The first parameter and the second parameter of the same sampled phase image are added together to obtain the cardiac motion parameters of the corresponding sampled phase.

[0092] In one embodiment, the equations for calculating the cardiac motion parameters of the sampled phase include:

[0093]

[0094] in, This is the average absolute difference between the current sampled phase image and the previous sampled phase image; The mean absolute difference between the current sampled phase image and the next sampled phase image. These are the cardiac motion parameters of the currently sampled phase image.

[0095] Step S406: Calculate the average optimal phase based on cardiac motion parameters from multiple phases.

[0096] Specifically, during cardiac systole, the phase with the minimum cardiac motion parameters among the sampling phases of cardiac systole is taken as the average optimal phase of cardiac systole; during cardiac diastole, the phase with the minimum cardiac motion parameters among the sampling phases of cardiac diastole is taken as the average optimal phase of cardiac diastole.

[0097] In one embodiment, the equation for calculating the mean optimal phase of cardiac systole includes:

[0098]

[0099] in, This represents the average optimal phase of cardiac systole; () represents the range of the sampling phase during cardiac systole.

[0100] In one embodiment, the equation for calculating the mean optimal phase of cardiac diastole includes:

[0101]

[0102] in, This represents the average optimal phase of cardiac diastole; () represents the range of the sampling phase during cardiac diastole.

[0103] The method described above for calculating the average optimal phase can calculate the cardiac motion parameters of the corresponding sampling phases based on the average absolute difference between each pair of sampling phases, and select the phase with the minimum cardiac motion parameters as the average optimal phase. This method can accurately determine the average optimal phase and ensure the accuracy of the optimal cardiac phase.

[0104] In one embodiment, such as Figure 5 As shown, a method for extracting a region of interest from an image is provided, including the following steps:

[0105] Step S502: Select a phase image within a preset range near the average optimal phase.

[0106] Specifically, the preset range is extended outwards from the average optimal phase by a certain number of pixels. Too small a phase range may miss the optimal phase, while too large a phase range may cause discontinuities in the optimal phase image sequence between different cardiac cycles. Therefore, the selection of the preset range is particularly important. In this embodiment, the preferred preset range is centered on the average optimal phase, extending forward by 10% and backward by 10%.

[0107] Step S504: Smooth the phase image within a preset range using a Gaussian low-pass filter.

[0108] Specifically, a Gaussian low-pass filter is used to smooth the phase image within a preset range. The Gaussian low-pass filter can eliminate the effects of noise, resulting in a smooth image that facilitates subsequent image processing.

[0109] Step S506: Extract the ventricular region image from the smoothed phase image.

[0110] Specifically, image segmentation is first performed based on the smoothed phase image and a bone threshold. Images with values ​​greater than the bone threshold are selected as the bone region image. Next, the bone region image undergoes maximum density projection along the thoracic cavity axis to obtain the maximum density projection image of the bone region. This maximum density projection is generated by calculating the highest density pixels encountered along each ray along the target area of ​​the patient. That is, when a ray passes through the smoothed phase image, the pixels with the highest density are retained and projected onto a two-dimensional plane, thus forming the maximum density projection image of the bone region. The thoracic cavity contour boundary is extracted from the maximum density projection image of the bone region. Based on the maximum density projection image of the bone region, different Boolean values ​​are assigned to the maximum density projection image of the bone region, and the thoracic cavity contour boundary is determined based on the boundaries of these different Boolean values. The ventricular region image is obtained based on the smoothed phase image and the thoracic cavity contour boundary. Images within the thoracic cavity contour boundary are selected from the smoothed phase image as the thoracic cavity image. Finally, the connected components of the thoracic cavity contour image are calculated, and the image within the connected component with the largest number of pixels is selected as the ventricular region image. A connected region is a region on the complex plane. If any simple closed curve is drawn in this region and the interior of the closed curve always belongs to this region, then this region is called a connected region.

[0111] Step S508: Calculate the contrast agent threshold based on the ventricular region image.

[0112] Specifically, region of interest extraction requires contrast agent thresholding for segmentation. Since the CT values ​​of different contrast agent concentrations vary, an empirical threshold cannot be used to segment contrast agent-containing regions. Therefore, a contrast agent threshold based on the current image and environment is needed. The gradient image of the ventricular region is calculated based on the ventricular region image. In image processing, the magnitude of the gradient is simply called the gradient, and the image composed of image gradients is called the gradient image. When edges exist in the image, there will always be a larger gradient value; conversely, when there are relatively smooth parts in the image, the gray value changes less, and the corresponding gradient is also smaller. Preferably, the Sobel operator is used to calculate the gradient image. The Sobel operator is a discrete first-order difference operator used to calculate an approximate value of the first-order gradient of the image brightness function. Applying this operator to any point in the image will generate the gradient vector corresponding to that point. The gray values ​​of all pixels in the gradient image are statistically analyzed, and ventricular region images with gradient image gray values ​​greater than a certain threshold at corresponding positions are used as marker images. Preferably, the gray values ​​of all pixels in the gradient image are statistically analyzed, and a histogram of all pixels is obtained. An appropriate proportion of gray values ​​is selected as a threshold, and gray values ​​greater than this threshold are selected as the labeled image. Then, the Otsu algorithm is used to calculate the contrast agent threshold for each pixel in the labeled image. The Otsu algorithm is an efficient algorithm for image binarization, using a threshold to divide the original image into foreground and background images. The optimal segmentation threshold obtained is used as the contrast agent threshold.

[0113] Step S510: Perform image segmentation based on the ventricular region image and the contrast agent threshold to obtain the contrast agent region image.

[0114] Specifically, image segmentation is performed using a contrast agent threshold, and images of the ventricular region that are larger than the contrast agent threshold are used as contrast agent region images.

[0115] Step S512: Select the region of interest image in the contrast agent region image.

[0116] Specifically, the right coronary artery is a clinically significant artery in terms of motion compared to other vessels. The motion of the right coronary artery in each phase can reflect the cardiac motion in that phase. In the contrast agent region images, images located in the upper left half of the ventricle, with fewer contrast agent pixels and lower CT values ​​outside the vessel, are selected as images of the vessel of interest.

[0117] The above method for extracting images of vessels of interest selects the ventricular region image based on the smoothed phase image, calculates the contrast agent threshold based on the ventricular region image, and uses the contrast agent threshold to segment the ventricular region image to obtain the contrast agent region image. Selecting the region of interest image from the contrast agent region image can accurately determine the right coronary vessel image within a preset range near the average optimal phase, further making the selection of the optimal phase of the heart more accurate.

[0118] In one embodiment, such as Figure 6 As shown, a method for extracting images of the ventricular region is provided, including the following steps:

[0119] Step S602: Perform image segmentation based on the smoothed phase image and the bone threshold to obtain the bone region image.

[0120] Specifically, based on the bone threshold, images with values ​​greater than the bone threshold are selected as bone region images. Clinically, the threshold for intrathoracic bone is generally 1500 HU. This means that regions with a smoothed phase image greater than 1500 HU can be selected as bone region images.

[0121] Step S604: Perform maximum density projection on the bone region image along the thoracic cavity axis to obtain the maximum density projection image of the bone region image.

[0122] Specifically, the maximum density projection is generated by calculating the highest density pixels encountered along each ray along the target area of ​​the patient. That is, when light passes through the initial optimal phase image sequence, the pixels with the highest density in the image are preserved and projected onto a two-dimensional plane, thus forming the maximum density projection image of the bone region.

[0123] Step S606: Calculate the thoracic cavity contour boundary based on the maximum density projection image of the bone region image.

[0124] Specifically, based on the maximum density projection image of the bone region image, the Boolean value of the central ventricle region in the maximum density projection image of the bone region image is set to 1, and the Boolean value of the non-ventricular region in the maximum density projection image of the bone region image is set to 0. The boundary between Boolean values ​​1 and 0 is used as the boundary of the thoracic cavity contour.

[0125] Step S608: Based on the smoothed phase image and the thoracic cavity contour boundary, obtain the ventricular region image.

[0126] Specifically, a chest cavity image is obtained based on the smoothed phase image and the chest cavity contour boundary. Connected regions are calculated based on the chest cavity image, and the image in the connected region with the most pixels is selected as the ventricular region image.

[0127] The thoracic cavity image is obtained based on the smoothed phase image and the thoracic cavity contour boundary. The image within the thoracic cavity contour boundary is selected from the smoothed phase image as the thoracic cavity image. That is, the region in the smoothed phase image that is greater than the soft tissue threshold and has a Boolean value equal to 1 is selected as the thoracic cavity image.

[0128] The connected components are calculated based on the chest cavity image, and the image within the connected component with the largest number of pixels is selected as the ventricular region image. Similarly, the connected components are calculated based on the chest cavity contour image, and the image within the connected component with the largest number of pixels is selected as the ventricular region image. Here, a connected component is a region on the complex plane. If any simple closed curve drawn within it always falls within this region, then this region is called a connected component.

[0129] The method described above for extracting ventricular region images involves segmenting a smoothed phase image based on a bone threshold to obtain a bone region image, followed by maximum density projection onto the bone region image. The maximum density projection of the bone region image is then calculated based on this projection. The chest cavity contour boundary is calculated, and images within this boundary are selected as the chest cavity image. Furthermore, the connected components of the chest cavity image are calculated, and the image with the most pixels within these components is selected as the ventricular image. This approach allows for more accurate calculation of the chest cavity contour boundary, leading to a more precise determination of the ventricular region and more accurate selection of the heart's location.

[0130] In one embodiment, such as Figure 7 As shown, a method for extracting the center line of blood vessels in an image of a region of interest is provided, including the following steps:

[0131] Step S702: Obtain coronal and sagittal view images of multiple regions of interest.

[0132] Specifically, the region of interest (ROI) image obtained in the above steps may contain multiple connected regions. These may include branches of blood vessels that are actually present on the current slice, or they may contain other non-vascular connected regions, such as calcification, metal, and bone, which may all be classified into the ROI. Therefore, for subsequent image processing, false positive vessels must first be eliminated, and the centerline of the vessels must be extracted. Since blood vessels are continuous in both the coronal and sagittal planes, it is necessary to first obtain images from both coronal and sagittal perspectives to eliminate false positive vessels. This involves acquiring coronal and sagittal perspective images of multiple ROI images. The coronal plane, also known as the frontal plane, is a cross-section that longitudinally divides the human body into anterior and posterior parts along its long axis from left to right. The sagittal plane is the anatomical plane that divides the human body into left and right sides; the plane parallel to the coronal plane is the sagittal plane.

[0133] Step S704: Determine the main vascular trunk based on the coronal and sagittal view images.

[0134] Specifically, the main vessel trunk is located in the center of the region of interest image and is the largest connected region. The main vessel trunk was determined based on coronal and sagittal view images.

[0135] Step S706: Filter out false positive blood vessels based on the main blood vessel trunk.

[0136] Specifically, based on the main vascular trunk, non-main trunk vessels are filtered out, and based on the main vascular trunk after filtering non-main trunk vessels, main trunk vessels are filtered out. False positive vessels are non-vascular regions. True positive vessels must meet two requirements: smoothness (the distance between the undetermined vessel and the determined vessel in the X-direction cannot be too large, and the distance between the undetermined vessel and the determined vessel in the cross-section cannot be too large; the position of the undetermined vessel is the location of the maximum value of the connected region of the non-main trunk, and the determined vessel is the location of the maximum value of the main trunk vessel closest to the undetermined vessel); and continuity (the distance between the undetermined vessel and the determined vessel on the y-axis cannot be too large, where the y-axis refers to the tomographic direction; when no connected region is detected in a certain tomographic section, the distance is 1; when no connected region is detected in multiple tomographic sections, continuity is not satisfied). When the undetermined vessel simultaneously satisfies both smoothness and continuity, the current undetermined vessel is considered a determined vessel. The mean of the x-direction positions of all determined vessels is used, and the undetermined vessel closest to this mean is considered a valid vessel. The mean of the x-direction positions of the determined vessels in the main trunk is used, and the undetermined vessel closest to this mean is considered a valid vessel.

[0137] Step S708: Determine the location of the vessel center in each segment based on the main vessel trunk after filtering out false positive vessels.

[0138] Specifically, based on the main vessels after filtering out false positive vessels, interpolation is used for filling, and the vessel center position of each slice is determined according to the vessel center positions in the sagittal and coronal planes.

[0139] Step S710: Based on the location of the blood vessel center in each slice, obtain the blood vessel centerline of the corresponding region of interest image.

[0140] The method described above for extracting the vessel centerline from the corresponding region of interest (ROI) image involves acquiring coronal and sagittal view images from the ROI image, identifying the main vessel trunks, filtering out false positive vessels based on the main vessel trunks, and further determining the vessel center position for each slice to obtain the vessel centerline of the corresponding ROI image. This method can more accurately identify the vessels of interest, making the selection of the heart location more precise.

[0141] In one embodiment, such as Figure 8 As shown, a method for extracting an image to be evaluated is provided, including the following steps:

[0142] Step S802: Perform a top-hat transform on the region of interest image to obtain a region of interest image that highlights the target object.

[0143] Specifically, top-hat transform is an image processing method that reduces the background in an image, making the target object stand out more. In this case, top-hat transform is applied to the region of interest (ROI) image to make the target object within the ROI image more prominent. The target object in this example is a blood vessel; by applying top-hat transform to the ROI image, the background is reduced, making the blood vessel appear clearer.

[0144] Step S804: Based on the soft tissue threshold, segment the region of interest image that preserves the intraventricular region.

[0145] Specifically, the empirical threshold for soft tissue is 800 HU. Based on the soft tissue threshold, the region of interest image of the prominent target object is segmented to obtain a region of interest image that preserves the intraventricular region.

[0146] Step S806: Select the region of interest image of the preserved intraventricular region within a preset range centered on the vascular centerline as the image to be evaluated for the corresponding phase.

[0147] Specifically, a region of interest (ROI) image centered on the vascular centerline and within a preset range preserving the intraventricular region is selected as the image to be evaluated for the corresponding phase. The preset range is defined as the region of interest within each slice, centered on the vascular centerline. Each pixel is used as the image to be evaluated for the current tomography. N is taken as a physical size between 50-100mm. Before segmentation, it is necessary to determine whether N has exceeded the boundary of the current image. For images that exceed the boundary, compensation is required before segmentation. The compensation value can be 0 or the boundary value. The image to be evaluated is extracted from the phase images within 10% of the average optimal phase, resulting in an image matrix of the diastolic and systolic phases.

[0148] The above-described method for extracting the image to be evaluated can accurately determine the range of the image to be evaluated based on the blood vessel centerline, further making the selection of the optimal phase more precise.

[0149] In one embodiment, such as Figure 9 As shown, a method for calculating the quality index of an image to be evaluated is provided, including the following steps:

[0150] Step S902: Based on the image to be evaluated and the segmentation threshold, obtain the image of the blood vessel of interest.

[0151] Specifically, before calculating the quality index, the resolution of the image to be evaluated can be increased. Increasing the resolution means improving the accuracy of calculating blood vessel morphology and edges. Two-dimensional image interpolation can be used to increase the resolution. The maximum gray value of the image to be evaluated is selected. The maximum gray value at multiple preset multiples is used as a segmentation threshold. Images of the image to be evaluated with gray values ​​greater than the segmentation threshold are designated as blood vessel images of interest for the corresponding segmentation threshold. Segmentation using multiple segmentation thresholds will yield multiple blood vessel images of interest. Preferably, three segmentation thresholds are obtained according to three preset multiples of the maximum gray value. The image to be evaluated is segmented using the first segmentation threshold, and the regions in the image to be evaluated with pixel gray values ​​greater than the first segmentation threshold are designated as the first blood vessel image of interest; the image to be evaluated is segmented using the second segmentation threshold, and the regions in the image to be evaluated with pixel gray values ​​greater than the second segmentation threshold are designated as the second blood vessel image of interest; the image to be evaluated is segmented using the third segmentation threshold, and the regions in the image to be evaluated with pixel gray values ​​greater than the third segmentation threshold are designated as the third blood vessel image of interest.

[0152] Step S904: Calculate the perimeter and area of ​​the target object in each image to be evaluated based on the image of the blood vessel of interest.

[0153] Specifically, based on the obtained multiple images of vessels of interest, the perimeter and area of ​​the target object in each image to be evaluated are calculated, that is, the perimeter and area of ​​the vessels in each image of vessels of interest are calculated.

[0154] Step S906: Calculate the degree of morphological regularity of the corresponding image to be evaluated based on the perimeter and area of ​​the target object in each image to be evaluated.

[0155] Step S908: Calculate the edge sharpness of the corresponding image to be evaluated based on the boundary of the image of the vessel of interest and the gradient map of the image of the vessel of interest.

[0156] Step S910: Calculate the quality index of each image to be evaluated based on the regularity of its shape and the sharpness of its edges.

[0157] Specifically, the number of vessels at the same location may differ across different phases. Considering that subsequent comparisons need to be made on the same baseline, i.e., the number of vessels in images at the same location across all phases must be consistent, a reference base number of vessels is introduced here.

[0158] Based on a comprehensive consideration of the number of basic blood vessels and the number of blood vessels in the image to be evaluated, a morphological regularity matrix and an edge sharpness matrix are obtained for each phase of the image to be evaluated. Since the magnitudes of morphological regularity and edge sharpness are inconsistent, it is necessary to bring the two metrics to the same baseline. This can be achieved through weighted averages or by normalizing the morphological regularity and edge sharpness. Finally, a quality index matrix is ​​generated for both diastole and systole. The cardiac cycle containing the quality index matrix is ​​identified. When the number of images containing the quality index matrix in that cardiac cycle exceeds the average, the quality indices corresponding to the images of each phase within each cardiac cycle are averaged, and the phase with the maximum value is selected as the optimal phase within that cardiac cycle. For cardiac cycles without a quality index matrix, the optimal phase of the nearest cardiac cycle containing a quality index matrix is ​​selected as its optimal phase. The optimal phase image of the entire image sequence can be extracted based on the optimal phase of each cardiac cycle. The optimal phase image of the corresponding cardiac cycle can be extracted from each reconstructed phase image and used as a synthetic sequence. Alternatively, the optimal phase of each phase can be used during the reconstruction of the projection domain to directly obtain the optimal phase image sequence.

[0159] The method described above for calculating the quality index can more accurately calculate the quality index of the phase image, thereby accurately determining the optimal phase of the heart and obtaining an image of the optimal phase of the heart based on the optimal phase.

[0160] The image reconstruction method described above does not rely on a human interface and can automatically detect and extract vessels of interest, as well as automatically analyze the quality of vessel images. In coronary angiography, it can be used to automatically select the optimal phase, eliminating the need for physicians to evaluate images and select reconstruction phases. This simplifies the coronary reconstruction process and saves physicians time in evaluating images and selecting parameters.

[0161] It should be understood that, although Figure 1-9 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1-9 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0162] In one embodiment, such as Figure 10As shown, an image reconstruction apparatus is provided, comprising: an image selection module 100 to be evaluated, a quality index calculation module 200, and an image reconstruction module 300, wherein:

[0163] The image selection module 100 is used to acquire the scan data of all phases and reconstruct multiple phase images as images to be evaluated.

[0164] The quality index calculation module 200 is used to calculate the quality index of each image to be evaluated according to the image quality evaluation rules.

[0165] The image reconstruction module 300 is used to calculate the optimal phase based on the quality index of each image to be evaluated, and to obtain the optimal phase image.

[0166] The image selection module 100 includes: an average optimal phase calculation unit, a region of interest image extraction unit, a blood vessel centerline extraction unit, and an image extraction unit for the image to be evaluated.

[0167] The average optimal phase calculation unit is used to calculate the average optimal phase based on the images corresponding to all phases.

[0168] The region of interest image extraction unit is used to select phase images within a preset range near the average optimal phase, and extract the region of interest image from the selected multiple phase images.

[0169] The vessel centerline extraction unit is used to extract the vessel centerline of the corresponding region of interest image based on multiple region of interest images.

[0170] The image extraction unit is used to segment the image within a preset range, centered on the blood vessel centerline, to obtain multiple images to be evaluated.

[0171] For specific limitations regarding the image reconstruction apparatus, please refer to the limitations on the image reconstruction method above, which will not be repeated here. Each module in the aforementioned image reconstruction apparatus can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0172] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 11As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements an image reconstruction method. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0173] Those skilled in the art will understand that Figure 11 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0174] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0175] Multiple images to be evaluated are acquired; based on the images to be evaluated and a segmentation threshold, an image of the vessel of interest is obtained; and image quality is evaluated based on the image of the vessel of interest.

[0176] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0177] Multiple phase images are reconstructed from the scan data of all phases and used as images to be evaluated. The quality index of each image to be evaluated is calculated according to the image quality evaluation rules. The best phase is calculated based on the quality index of each image to be evaluated, and the best phase image is obtained.

[0178] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0179] Acquire scan data for all phases and reconstruct images corresponding to all phases. Calculate the average optimal phase based on the images corresponding to all phases. Select phase images within a preset range near the average optimal phase and extract regions of interest (ROIs) from the selected phase images. Extract the vessel centerline from the multiple ROI images. Segment the image within a preset range centered on the vessel centerline to obtain multiple images to be evaluated. Calculate the quality index for each image to be evaluated according to image quality evaluation rules. Calculate the optimal phase based on the quality index of each image to be evaluated and reconstruct the optimal phase image.

[0180] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0181] Calculate the mean absolute difference between two adjacent phase images based on their pixel values ​​and the image matrix size. Then, calculate the motion parameters for each phase based on the mean absolute difference. Finally, calculate the average optimal phase based on the cardiac motion parameters from multiple phases.

[0182] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0183] Phase images within a preset range near the average optimal phase are selected. These phase images within the preset range are smoothed using a Gaussian low-pass filter. The ventricular region image is extracted from the smoothed phase image. The contrast agent threshold is calculated based on the ventricular region image. Image segmentation is performed based on the ventricular region image and the contrast agent threshold to obtain the contrast agent region image. The region of interest is selected within the contrast agent region image.

[0184] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0185] Image segmentation is performed based on the smoothed phase image and bone thresholding to obtain the bone region image. Maximum density projection is then performed on the bone region image along the thoracic cavity axis to obtain the maximum density projection image of the bone region image. The thoracic cavity contour boundary is calculated based on the maximum density projection image of the bone region image. Finally, the ventricular region image is obtained based on the smoothed phase image and the thoracic cavity contour boundary.

[0186] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0187] Acquire coronal and sagittal images of multiple regions of interest (ROIs). Determine the main vascular trunks based on the coronal and sagittal images. Filter out false positive vessels based on the main vascular trunks. Determine the vessel center position for each slice based on the main vascular trunks after filtering out false positive vessels. Obtain the vessel centerline of the corresponding ROI image based on the vessel center position of each slice.

[0188] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0189] A top-hat transform is performed on the region of interest (ROI) image to obtain an ROI image highlighting the target object. Based on a soft tissue threshold, ROI images preserving the ventricular regions are segmented. ROI images preserving the ventricular regions within a preset range centered on the vessel centerline are selected as the evaluation images for the corresponding phases.

[0190] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0191] A preset multiple of the maximum grayscale value in the image to be evaluated is selected as the segmentation threshold. Based on the image to be evaluated and the segmentation threshold, the image of the vessel of interest is obtained. Based on the image of the vessel of interest, the perimeter and area of ​​the target object in each image to be evaluated are calculated. Based on the perimeter and area of ​​the target object in each image to be evaluated, the morphological regularity of the corresponding image to be evaluated is calculated. Based on the boundary and gradient map of the image of the vessel of interest, the edge sharpness of the corresponding image to be evaluated is calculated. Based on the morphological regularity and edge sharpness of the images to be evaluated, the quality index of each image to be evaluated is calculated.

[0192] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0193] Multiple images to be evaluated are acquired; based on the images to be evaluated and a segmentation threshold, an image of the vessel of interest is obtained; and image quality is evaluated based on the image of the vessel of interest.

[0194] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0195] Multiple phase images are reconstructed from the scan data of all phases and used as images to be evaluated. The quality index of each image to be evaluated is calculated according to the image quality evaluation rules. The best phase is calculated based on the quality index of each image to be evaluated, and the best phase image is obtained.

[0196] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0197] Acquire scan data for all phases and reconstruct images corresponding to all phases. Calculate the average optimal phase based on the images corresponding to all phases. Select phase images within a preset range near the average optimal phase and extract regions of interest (ROIs) from the selected phase images. Extract the vessel centerline from the multiple ROI images. Segment the image within a preset range centered on the vessel centerline to obtain multiple images to be evaluated. Calculate the quality index for each image to be evaluated according to image quality evaluation rules. Calculate the optimal phase based on the quality index of each image to be evaluated and reconstruct the optimal phase image.

[0198] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0199] Calculate the mean absolute difference between two adjacent phase images based on their pixel values ​​and the image matrix size. Then, calculate the motion parameters for each phase based on the mean absolute difference. Finally, calculate the average optimal phase based on the cardiac motion parameters from multiple phases.

[0200] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0201] Phase images within a preset range near the average optimal phase are selected. These phase images within the preset range are smoothed using a Gaussian low-pass filter. The ventricular region image is extracted from the smoothed phase image. The contrast agent threshold is calculated based on the ventricular region image. Image segmentation is performed based on the ventricular region image and the contrast agent threshold to obtain the contrast agent region image. The region of interest is selected within the contrast agent region image.

[0202] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0203] Image segmentation is performed based on the smoothed phase image and bone thresholding to obtain the bone region image. Maximum density projection is then performed on the bone region image along the thoracic cavity axis to obtain the maximum density projection image of the bone region image. The thoracic cavity contour boundary is calculated based on the maximum density projection image of the bone region image. Finally, the ventricular region image is obtained based on the smoothed phase image and the thoracic cavity contour boundary.

[0204] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0205] Acquire coronal and sagittal images of multiple regions of interest (ROIs). Determine the main vascular trunks based on the coronal and sagittal images. Filter out false positive vessels based on the main vascular trunks. Determine the vessel center position for each slice based on the main vascular trunks after filtering out false positive vessels. Obtain the vessel centerline of the corresponding ROI image based on the vessel center position of each slice.

[0206] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0207] A top-hat transform is performed on the region of interest (ROI) image to obtain an ROI image highlighting the target object. Based on a soft tissue threshold, ROI images preserving the ventricular regions are segmented. ROI images preserving the ventricular regions within a preset range centered on the vessel centerline are selected as the evaluation images for the corresponding phases.

[0208] In one embodiment, when the computer program is executed by a processor, it also performs the following steps:

[0209] A preset multiple of the maximum grayscale value in the image to be evaluated is selected as the segmentation threshold. Based on the image to be evaluated and the segmentation threshold, the image of the vessel of interest is obtained. Based on the image of the vessel of interest, the perimeter and area of ​​the target object in each image to be evaluated are calculated. Based on the perimeter and area of ​​the target object in each image to be evaluated, the morphological regularity of the corresponding image to be evaluated is calculated. Based on the boundary and gradient map of the image of the vessel of interest, the edge sharpness of the corresponding image to be evaluated is calculated. Based on the morphological regularity and edge sharpness of the images to be evaluated, the quality index of each image to be evaluated is calculated.

[0210] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0211] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

[0212] The above embodiments merely illustrate several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method of image reconstruction, characterized by, The method comprises: reconstructing all the images of multiple phases from scanning data to obtain images to be evaluated; automatically calculating a quality index of each image to be evaluated according to an image quality evaluation rule; selecting a best phase according to the quality index of each image to be evaluated, and reconstructing a best phase image according to the best phase.

2. The method of claim 1, wherein, The automatically calculating a quality index of each image to be evaluated according to an image quality evaluation rule comprises: extracting blood vessels from each image to be evaluated to obtain multiple images of blood vessels of interest; calculating a quality index of each image to be evaluated according to the multiple images of blood vessels of interest.

3. The method of claim 2, wherein, The calculating a quality index of each image to be evaluated according to the multiple images of blood vessels of interest comprises: calculating the perimeter and area of the blood vessels in each image of blood vessels of interest to obtain a morphological regularity degree of the corresponding image to be evaluated; obtaining an edge sharpness degree of the corresponding image to be evaluated according to the boundary of each image of blood vessels of interest and a gradient map of the image of blood vessels of interest; calculating a quality index of each image to be evaluated according to the morphological regularity degree and the edge sharpness degree.

4. The method of claim 1, wherein, The method further comprises: obtaining the number of basic blood vessels and the number of blood vessels in the image to be evaluated; obtaining a morphological regularity degree matrix and an edge sharpness degree matrix of each phase of the image to be evaluated according to the number of basic blood vessels and the number of blood vessels; normalizing or weighting the morphological regularity degree matrix and the edge sharpness degree matrix to obtain a quality index matrix; determining a cardiac cycle in which the quality index matrix is located, and selecting a phase corresponding to the average maximum value of the quality index of each phase in each cardiac cycle as the best phase in the cardiac cycle when the number of images of the quality index matrix contained in the cardiac cycle exceeds the average number; selecting the best phase of the cardiac cycle containing the quality index matrix adjacent to the cardiac cycle without the quality index matrix as the best phase.

5. The method of claim 1, wherein, The selecting a best phase according to the quality index of each image to be evaluated, and reconstructing a best phase image according to the best phase comprises: automatically selecting an image to be evaluated with the maximum quality index according to the quality index of each image to be evaluated; selecting the phase of the image to be evaluated with the maximum quality index as the best phase, and reconstructing a best phase image according to the best phase.

6. The method according to any one of claims 1 to 5, characterized in that, The reconstructing all the images of multiple phases from scanning data to obtain images to be evaluated comprises: calculating an average best phase according to the images corresponding to multiple phases; selecting phase images within a preset range near the average best phase, and extracting images of regions of interest from the selected multiple phase images; extracting a blood vessel center line of the corresponding image of a region of interest according to multiple images of regions of interest; performing image segmentation within a preset range with the blood vessel center line as the center to obtain multiple images to be evaluated.

7. The method of claim 6, wherein, The selecting phase images within a preset range near the average best phase, and extracting images of regions of interest from the selected multiple phase images comprises: selecting a phase image in a preset range around the average optimal phase; smoothing the phase image in the preset range by using a Gaussian low-pass filter; extracting a ventricular region image from the smoothed phase image; calculating a contrast agent threshold according to the ventricular region image; performing image segmentation according to the ventricular region image and the contrast agent threshold to obtain a contrast agent region image; selecting a region of interest image in the contrast agent region image.

8. The method of claim 6, wherein, The extracting a blood vessel centerline of the corresponding region of interest image according to the multiple region of interest images comprises: obtaining a coronal view image and a sagittal view image of the multiple region of interest images; determining a blood vessel trunk according to the coronal view image and the sagittal view image; filtering false positive blood vessels according to the blood vessel trunk; determining a blood vessel center position of each slice according to the blood vessel trunk after filtering the false positive blood vessels; obtaining a blood vessel centerline of the corresponding region of interest image according to the blood vessel center position of each slice.

9. The method of claim 6, wherein, The image segmentation in a preset range with the blood vessel centerline as the center to obtain multiple to-be-evaluated images comprises: performing top-hat transformation on the region of interest image to obtain a region of interest image highlighting target objects; segmenting to obtain a region of interest image retaining an intraventricular region according to a soft tissue threshold; selecting the region of interest image retaining the intraventricular region in the preset range with the blood vessel centerline as the center as a to-be-evaluated image of a corresponding phase.

10. An image reconstruction apparatus, characterized by comprising: The device comprises: a to-be-evaluated image selection module configured to obtain all multiple phase images reconstructed from scanning data as to-be-evaluated images; a quality index calculation module configured to automatically calculate a quality index of each to-be-evaluated image according to an image quality evaluation rule; an image reconstruction module configured to select a best phase according to the quality index of each to-be-evaluated image, and to reconstruct a best phase image according to the best phase.

Images