CT image quality method and system for improving micro-radiation dose

Abstract

The invention relates to the technical field of image data processing, in particular to a CT image quality method and system for improving micro-radiation dose, and the method comprises the steps: obtaining a plurality of continuous frames of CT scanning images of the same part of a detected person, screening out a plurality of to-be-selected matching points from the feature points of each frame of CT scanning image, screening out a plurality of key points from the to-be-selected matching points, obtaining the motion vector of the key points in each frame of CT scanning image according to the matching of the key points in the adjacent frames of CT scanning images, and embedding the motion vectors of the key points in all frames of CT scanning images into a statistical iteration reconstruction algorithm as priori knowledge. And outputting a reconstructed CT scanning image through the updated statistical iteration reconstruction algorithm. According to the method, the accurate motion vector of the key point in the CT scanning image is obtained, so that the reconstruction quality of the CT image of the micro-radiation dose is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method and system for improving CT image quality by micro-radiation dose. Background Art

[0002] Micro-radiation CT imaging refers to a new type of CT imaging technology that uses various imaging algorithms and technologies to achieve extremely low radiation doses while maintaining or even improving image quality. It can significantly reduce the radiation dose during CT examinations while ensuring image quality, thereby reducing radiation damage to patients. The implementation of micro-radiation CT mainly relies on image reconstruction algorithms, which first construct an initial scan image, integrate the initial image at different angles to obtain a simulated projection image, and then compare the difference between the simulated projection image and the actual projection image. This difference is then back-propagated to adjust the initial image, and this process is repeated continuously to continuously reduce the error, ultimately obtaining a micro-radiation CT image.

[0003] However, when scanning the human body, various abdominal organs are constantly moving (such as intestinal peristalsis and heartbeats). This movement can lead to motion artifacts. This requires constructing motion vectors by taking the differences between corresponding feature points across multiple images. These motion vectors are then used to compensate for the movement of these organs and eliminate motion artifacts.

[0004] Existing Problems: Existing feature point selection and matching methods primarily rely on optical flow and deformation registration. These methods use corner detection algorithms, such as the SIFT (Scale-Invariant Feature Transform) algorithm, to determine feature points at different times. Feature points between adjacent images are then matched using techniques such as nearest neighbor search and block matching. However, while micro-radiation CT images can be more accurately positioned during scanning, projection magnifies motion artifacts in the projected image, making it difficult to accurately locate the matching relationships between identified feature points. This results in errors in the resulting motion vector, making it impossible to eliminate the intended artifacts and potentially leading to poor quality micro-radiation CT images. Summary of the Invention

[0005] The present invention provides a method and system for improving CT image quality by micro-radiation dose, so as to solve the existing problems.

[0006] The present invention provides a method and system for improving CT image quality with micro-radiation doses using the following technical solutions: An embodiment of the present invention provides a method for improving the quality of CT images with micro-radiation doses, the method comprising the following steps: Obtaining a number of consecutive frames of CT scan images of the same part of the person being tested; Taking any frame of CT scan image as the target image, several frames of images with the same frequency are selected from the CT scan images other than the target image; obtaining several feature points in each frame of CT scan image, and selecting several candidate matching points from the feature points in the target image; According to the matching conditions of the candidate matching points in the target image and the same-frequency image, a number of key points are selected from the candidate matching points in the target image; According to the matching results of key points in adjacent frames of CT scan images, the motion vectors of the key points in each frame of CT scan image are obtained; the motion vectors of the key points in all frames of CT scan image are embedded as prior knowledge into the statistical iterative reconstruction algorithm to obtain an updated statistical iterative reconstruction algorithm; each frame of CT scan image is input into the updated statistical iterative reconstruction algorithm, and a reconstructed CT scan image is output.

[0007] Furthermore, the step of selecting a plurality of frames of same-frequency images from the CT scan images other than the target image includes the following specific steps: In any row of any frame of CT scan image, the grayscale values of all pixels are counted from left to right to form a grayscale value sequence; All frames of CT scan images except the target image are recorded as reference images; The target image is compared with the The average of the inversely proportional normalized values of the DTW distances of the grayscale value sequences of all the same rows of the frame reference image is recorded as the distance between the target image and the first The overall image similarity of the frame reference image; According to the similarity between the target image and each frame of the reference image, the same-frequency image of the target image is determined.

[0008] Furthermore, the step of determining the same-frequency image of the target image according to the overall image similarity between the target image and each frame of the reference image includes the following specific steps: The reference image whose overall image similarity is greater than or equal to the preset similarity threshold is recorded as the same-frequency image of the target image.

[0009] Furthermore, the specific steps of selecting a plurality of candidate matching points from the feature points in the target image include the following: For any feature point in the target image , in Acquisition and feature points in the same frequency image The feature point with the closest Euclidean distance to the position coordinates is recorded as the feature point Matching point, if the feature point With feature points When the Euclidean distance of the position coordinates of the matching point is less than the preset distance threshold, the feature point Recorded as the matching point to be selected.

[0010] Furthermore, the specific steps of selecting a plurality of key points from the candidate matching points in the target image include the following: In the target image, each candidate matching point is used as the initial cluster center. According to the Euclidean distance between pixels, all pixels are clustered only once to obtain the connected area corresponding to each candidate matching point. For any connected region , to connect the regions The center point to the connected area The Euclidean distance and direction of each adjacent connected area center point are the vectors, which are recorded as connected areas. The corresponding neighborhood vector of the candidate matching point; Obtain the neighborhood vector of each candidate matching point in each frame of the same-frequency image according to the method of obtaining the neighborhood vector of each candidate matching point in the target image; According to the target image The first of the candidate matching points The neighborhood vector and the The candidate matching points are The cosine similarity of each neighborhood vector of the matching point in the same-frequency image of the frame is used to determine the first The first of the candidate matching points The maximum similarity of neighborhood vectors; According to the maximum similarity, the first Filter out matching vectors from the neighborhood vectors of the candidate matching points; The target image The mean of the maximum similarity of all matching vectors of the candidate matching points is recorded as The candidate matching points are Structural matching in frame-same frequency images; According to the structural matching degree of each candidate matching point in the target image in each frame of the same-frequency image, several key points are screened out.

[0011] Furthermore, the determination of the first The first of the candidate matching points The maximum similarity of neighborhood vectors includes the following specific steps: The target image The first of the candidate matching points The neighborhood vector and the The candidate matching points are The maximum value of the cosine similarity of all neighborhood vectors of the matching point in the same-frequency image is recorded as the first The first of the candidate matching points The maximum similarity of the neighborhood vectors.

[0012] Furthermore, according to the maximum similarity, the first The matching vectors are screened out from the neighborhood vectors of the candidate matching points. The specific steps are as follows: For the target image Among all the neighborhood vectors of the candidate matching points, the neighborhood vector with the maximum similarity greater than or equal to the preset vector threshold is recorded as the matching vector.

[0013] Furthermore, the method of selecting a plurality of key points based on the structural matching degree of each candidate matching point in the target image in each frame of the same-frequency image includes the following specific steps: The product of the preset quantity coefficient and the number of images with the same frequency is recorded as the quantity threshold; For the target image The structural matching degree of the candidate matching points in all frames of the same-frequency images is calculated, and the same-frequency images with a structural matching degree greater than or equal to the preset matching threshold are recorded as matching same-frequency images. When the number of matching same-frequency images is greater than the number threshold, the first image in the target image is recorded as the matching same-frequency image. The matching points to be selected are recorded as key points.

[0014] Furthermore, the step of obtaining the motion vector of the key point in each frame of the CT scan image includes the following specific steps: By brute force matching, get the Each key point of the frame CT scan image is The key points of successful matching on the frame CT scan image; Get the CT scan image frame The position coordinates of the key points to the CT scan image frame The key point is The vector formed by the Euclidean distance and direction of the position coordinates of the key points that are successfully matched on the frame CT scan image is recorded as CT scan image frame The motion vector of each key point.

[0015] The present invention also proposes a system for improving CT image quality with micro-radiation dose, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned method for improving CT image quality with micro-radiation dose.

[0016] The beneficial effects of the technical solution of the present invention are: In an embodiment of the present invention, several consecutive frames of CT scan images of the same part of the person being examined are acquired, and several candidate matching points are selected from the feature points in each frame of the CT scan image. Furthermore, several key points are selected from the candidate matching points in each frame of the CT scan image. Thus, based on the rhythmic patterns of organ tissue movement, key points that can maintain their original structure during motion are selected, thereby improving the accuracy of key point selection and ensuring the accuracy of subsequent motion vector analysis of key points. Based on the key point matching in adjacent frames of the CT scan image, the motion vector of the key point in each frame of the CT scan image is acquired. The motion vectors of the key points in all frames of the CT scan image are embedded as prior knowledge into a statistical iterative reconstruction algorithm to obtain an updated statistical iterative reconstruction algorithm. The statistical iterative reconstruction algorithm is then motion-corrected using the motion vectors, ensuring high-quality reconstruction of subsequent images. Each frame of the CT scan image is input into the updated statistical iterative reconstruction algorithm, which outputs a reconstructed CT scan image. At this point, the present invention obtains the precise motion vectors of key points in CT scan images, that is, combines the structural characteristics of organs to effectively avoid the problem of feature point matching errors caused by texture similarity, thereby greatly improving the efficiency of eliminating motion artifacts in CT scan images and improving the quality of CT image reconstruction with micro-radiation doses. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 This is a flowchart of the steps of a method for improving CT image quality with micro-radiation dose according to the present invention; Figure 2 This is a schematic diagram of the CT machine structure; Figure 3 This is the reconstructed CT image without removing motion artifacts; Figure 4 CT images reconstructed after removing motion artifacts. DETAILED DESCRIPTION

[0019] To further illustrate the technical means and effectiveness of the present invention in achieving its intended objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effectiveness of a method and system for improving CT image quality with micro-radiation doses according to the present invention. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0020] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0021] The following describes in detail a method and system for improving CT image quality with micro-radiation dose provided by the present invention with reference to the accompanying drawings.

[0022] See also Figure 1 , which shows a flowchart of a method for improving CT image quality of micro-radiation dose provided by an embodiment of the present invention, the method comprising the following steps: Step S001: Acquire a plurality of consecutive frames of CT scan images of the same part of a person to be examined.

[0023] In this embodiment, a CT machine is used to collect a plurality of consecutive frames of CT scan images of the same part of a person for inspection. CT scan images are grayscale images.

[0024] The specific acquisition process is as follows: first, power on the entire system, use the self-test program to check the current status of each hardware component, and perform routine calibration and calibration on the CT machine. Then use the laser positioning system at the scan head to locate the organ to be inspected, input the exposure parameters of the radiation source, the acquisition parameters of the array detector, and the operating parameters of each axis motion mechanism, etc. Afterwards, according to the setting of the scanning parameters, the corresponding scanning movement is performed. Synchronously, the formed X-ray beam irradiates the human body and interacts with human tissue. The energy-changed light quanta enters a specific metal grid along a set direction and reaches the corresponding position of the two-dimensional array detector to be counted, and the corresponding two-dimensional array is obtained. These two-dimensional arrays are arranged to form a two-dimensional image of the CT scan. After the scan is completed, the motion mechanisms of each axis return to their positions, and the person being inspected is asked to get out of bed. Schematic diagram of the CT machine structure, as shown below Figure 2 As shown. The CT image reconstructed without removing motion artifacts is shown in Figure 3 shown.

[0025] It should be noted that during a CT scan, the patient lies on a bed, which is moved to a predetermined position. The scanner then releases a micro-dose of X-rays. These X-rays penetrate the body, interact with internal tissues, and are then detected by a special grid in the detector, forming a two-dimensional scan image of the corresponding position. During the scan, due to the constant movement of various tissues within the body (such as gastrointestinal peristalsis), the X-rays blur the interiors of organs and their edges. While the edges of different organs naturally have distinct boundary features, motion artifacts and the tilt of the X-rays as they irradiate the tissues cause the edges of the organs and tissues in the two-dimensional scan image to widen, compromising the accuracy of the original positioning based on feature points. However, since the rhythm of human organ tissues has certain rules, there will be a certain periodicity. Although the edge of the organ becomes wider due to movement and oblique illumination, when the organ tissue moves to the same position, since it is the image of the same organ, the common repeated structural features in multiple frames will be more obvious, which can more accurately reflect the original characteristics of the organ edge. Based on this, the feature points and feature edges in each scanned image can be constructed, and then the changes in these feature points between consecutive moments can be examined to construct the motion vector of each pixel.

[0026] Step S002: Taking any frame of CT scan image as the target image, a number of frames of images with the same frequency are selected from CT scan images other than the target image; obtaining a number of feature points in each frame of CT scan image, and selecting a number of candidate matching points from the feature points in the target image.

[0027] Preferably, in one embodiment of the present invention, the method for obtaining the candidate matching points includes: The SIFT algorithm is used to obtain several feature points in each frame of the CT scan image.

[0028] It should be noted that the SIFT (Scale-Invariant Feature Transform) algorithm is a well-known technique, and its specific method will not be described here. Regardless of how the organ in question moves, its surface will always have some wrinkles or curvatures. Although their positions may change due to movement, their characteristic structures remain constant. For example, where two organs meet, there will always be a certain intersection. Therefore, the SIFT algorithm can be used to process multiple scanned 2D images to obtain feature corner points in the scanned images. Specifically, the scanned images are first Gaussian blurred to construct a scale space and generate a corresponding differential Gaussian pyramid. Then, at each layer, the grayscale value information of the current pixel is compared with that of its neighbors, marking local extreme points. The Hessian matrix is then used to calculate the principal curvature ratio of these extreme points. If the ratio exceeds a threshold (preset to 10:1 in this example), it is determined to be an edge artifact (such as a stripe artifact at the edge of a rib) and is removed. Selectable feature points in the scanned 2D image are thus obtained.

[0029] It should be further explained that during a CT scan, human organs are constantly moving, and there are always certain patterns in the movement of human organs. After a certain period of time, they will return to their original positions. However, due to the large number of organs in the human abdomen, different organ tissues may be covered in the same cross-section (the target position of the CT scan). Therefore, it is necessary to find some images separated by one cycle based on the positions where the organs move to. This is to eliminate some candidate points (optional feature points) with large differences.

[0030] In any row of any frame of CT scan image, the grayscale values of all pixels are counted from left to right to form a grayscale value sequence.

[0031] Any frame of CT scan image is recorded as the target image, and all frames of CT scan image except the target image are recorded as reference images.

[0032] Get the target image and DTW distance of the gray value sequence in the same row of the frame reference image The inverse proportional normalization value of the target image is compared with the The average of the inversely proportional normalized values of the DTW distances of the grayscale value sequences of all the same rows of the frame reference image is recorded as the distance between the target image and the first The overall graph similarity of the frame reference image.

[0033] It should be noted that: in this embodiment, the DTW (Dynamic Time Warping) algorithm is used to obtain the DTW distance of two gray value sequences. The smaller the DTW distance, the more similar the two gray value sequences are. This is a well-known technology and the specific method will not be introduced here. To present The inverse proportional relationship and normalization processing can be set by the implementer according to the actual situation.

[0034] According to the above method, the overall image similarity between the target image and each frame of the reference image can be obtained.

[0035] The preset similarity threshold is 0.75, and the preset distance threshold is 5, which is used as an example for description.

[0036] The reference image whose overall image similarity is greater than or equal to the preset similarity threshold is recorded as the same-frequency image of the target image.

[0037] It should be noted that: when the similarity of the overall image is greater than or equal to the preset similarity threshold, it is considered that when the target image and the reference image are scanned, the organ tissue just moves to the same close position, that is, the same-frequency image, thereby obtaining several frames of the target image with the same frequency.

[0038] For any feature point in the target image , in Acquisition and feature points in the same frequency image The feature point with the closest Euclidean distance to the position coordinates is recorded as the feature point Matching point, if the feature point With feature points When the Euclidean distance of the position coordinates of the matching point is less than the preset distance threshold, the feature point Recorded as the matching point to be selected.

[0039] It should be noted that in this embodiment, a rectangular coordinate system is constructed with the vertex at the lower left corner of each frame of the CT scan image as the origin, the horizontal axis is the square to the right, and the vertical axis is the positive direction. The position coordinates of the feature points in each frame of the CT scan image are obtained in the rectangular coordinate system. Acquisition and feature points in the same frequency image When there are multiple feature points with the closest Euclidean distance to the position coordinates, any one of them is selected as the feature point If the feature point With feature points When the Euclidean distance of the position coordinates of the matching point is greater than or equal to the preset distance threshold, it means that the feature point If a large change occurs at the same moment in two cycles of tissue movement, the possibility that the two feature points are the same point on the organ is small, and it will be difficult to match them as matching points. They will be recorded as discarded points, otherwise they will be recorded as candidate matching points.

[0040] According to the above method, a plurality of candidate matching points in the target image and a matching point of each candidate matching point in each frame of the same-frequency image can be obtained.

[0041] Step S003: According to the matching conditions of the candidate matching points in the target image and the same-frequency image, a number of key points are screened out from the candidate matching points in the target image.

[0042] It should be noted that when tissues and organs move, each has its own unique structure and its position is fixed, meaning that adjacent organs are fixed. This results in unique texture features in the scanned images of different organs. Furthermore, during organ movement, because many organs are in motion, some organs may experience some deformation. This necessitates selecting unchanging feature points from the numerous available feature points. These feature points should maintain a relatively fixed angular structure regardless of movement, allowing for better matching of feature points at the same location in the two scanned images.

[0043] Preferably, in one embodiment of the present invention, the method for obtaining key points includes: In the target image, each candidate matching point is taken as the initial cluster center, and the Euclidean distance between the position coordinates of the pixel points is taken as the cluster distance. The K-means clustering algorithm is used to perform cluster assignment on all pixels only once, and each initial cluster center corresponds to a connected area, that is, each candidate matching point corresponds to a connected area, and there is only one candidate matching point in each connected area.

[0044] It should be noted that: since many candidate matching points are distributed throughout different organs and tissues, it is difficult to describe their adjacent relationships through fixed positional relationships, so each candidate matching point is used as the initial cluster center for single cluster assignment. Among them, the K-means clustering algorithm is a well-known technology, and the single cluster assignment is: for each pixel in the target image, calculate its Euclidean distance with all initial cluster centers, and assign the pixel to the nearest initial cluster center. In this way, each initial cluster center will correspond to an area, that is, an area composed of all pixels assigned to the center. And because the clustering process will cluster together pixels that are close in space, a continuous area is formed in the image, that is, one initial cluster center corresponds to one connected area.

[0045] According to the method of obtaining the connected area corresponding to each candidate matching point in the target image, the connected area corresponding to the matching point of each candidate matching point in each frame of the same-frequency image is obtained.

[0046] It should be noted that because organs move to different positions in different scanned images, different scanned images have different numbers of candidate matching points when detecting feature points and screening candidate matching points. To better describe the location information and surrounding structural information of each candidate feature point, it is necessary to describe the structure near the candidate matching point based on the relative position of the segmented area it is located in and the surrounding segmented areas.

[0047] In the target image, for any connected region , to connect the regions The center point to the connected area The Euclidean distance and direction of each adjacent connected area center point are the vectors, which are recorded as connected areas. The corresponding neighborhood vector of the candidate matching point. Thus, several neighborhood vectors of each candidate matching point are obtained.

[0048] According to the method of obtaining the neighborhood vector of each candidate matching point in the target image, the neighborhood vector of the matching point of each candidate matching point in each frame of the same-frequency image is obtained.

[0049] It should be noted that if the candidate matching point and its corresponding matching point in the target image and the same-frequency image have similar neighborhood features, that is, in certain general directions, there are one or more feature points at the same or similar locations, and the texture structures at these feature points are also similar, then it can be said that the candidate matching point and its corresponding matching point are at the same location on the organ. Based on this, the structural matching degree between the candidate matching point and its corresponding matching point in the target image and the same-frequency image is constructed.

[0050] Calculate the target image The first of the candidate matching points The neighborhood vector and the The candidate matching points are The cosine similarity of each neighborhood vector of the matching point in the same-frequency image of the frame is used to calculate the first The first of the candidate matching points The neighborhood vector and the The candidate matching points are The maximum value of the cosine similarity of all neighborhood vectors of the matching point in the same-frequency image is recorded as the first The first of the candidate matching points The maximum similarity of the neighborhood vectors.

[0051] It should be noted that the calculation of the cosine similarity between two vectors is a well-known technique and the specific method will not be introduced here. The value range of the cosine similarity is between -1 and 1, where a cosine similarity of 1 indicates that the two vectors are exactly the same.

[0052] According to the above method, the maximum similarity of each neighborhood vector of each candidate matching point in the target image can be obtained.

[0053] The preset vector threshold is 0.5, the preset matching threshold is 0.7, and the preset quantity coefficient is 50%, which is used as an example for description.

[0054] For the target image Among all the neighborhood vectors of the candidate matching points, the neighborhood vector with the maximum similarity greater than or equal to the preset vector threshold is recorded as the matching vector.

[0055] The target image The mean of the maximum similarity of all matching vectors of the candidate matching points is recorded as The candidate matching points are Structural matching in frame-by-frame images.

[0056] The product of the preset quantity coefficient and the number of images with the same frequency is recorded as the quantity threshold.

[0057] For the target image The structural matching degree of the candidate matching points in all frames of the same-frequency images is calculated, and the same-frequency images with a structural matching degree greater than or equal to the preset matching threshold are recorded as matching same-frequency images. When the number of matching same-frequency images is greater than the number threshold, the first image in the target image is recorded as the matching same-frequency image. The matching points to be selected are recorded as key points.

[0058] According to the above method, several key points in the target image can be obtained.

[0059] What needs to be explained is: when the target image The greater the structural matching degree of the candidate matching point in each frame of the same frequency image, the greater the The greater the possibility that the candidate matching point and its matching point in each frame of the same-frequency image are the same point on the organ, the matching same-frequency images can be screened. When the number of matching same-frequency images exceeds half of the number of same-frequency images, it means that the structural matching degree of the candidate matching point in more same-frequency images is greater, and the candidate matching point is more critical.

[0060] According to the above method, several key points in each frame of CT scan image can be obtained.

[0061] Step S004: Based on the matching results of key points in adjacent frames of CT scan images, obtain the motion vector of the key points in each frame of the CT scan image; embed the motion vector of the key points in all frames of the CT scan image as prior knowledge into the statistical iterative reconstruction algorithm to obtain an updated statistical iterative reconstruction algorithm; input each frame of the CT scan image into the updated statistical iterative reconstruction algorithm, and output a reconstructed CT scan image.

[0062] Preferably, in one embodiment of the present invention, the method for obtaining the reconstructed CT scan image includes: For adjacent frames of CT scan images, the Brute-Force method is used to obtain the first Each key point of the frame CT scan image is Key points that are successfully matched on the frame CT scan image.

[0063] It should be noted that Brute-Force is a basic feature matching method, which is a well-known technology and the specific method will not be introduced here. As mentioned above, many key points are screened out in each scanned image. These key points can better describe the position information of the organ at the corresponding moment. Therefore, in this embodiment, Brute-Force is performed on the two scanned images, and the grid division method (well-known technology) is combined to perform spatial constraints to accelerate the matching speed. That is, all the contrast nodes in the grid divided by the scanned images of two adjacent frames are traversed, and the Euclidean distance between the descriptors corresponding to the contrast nodes is calculated (the contrast key points are converted into high-dimensional gradient histogram descriptors based on the Gaussian difference pyramid), and the Euclidean distance between the many descriptors is obtained. If the Euclidean distance between the two descriptors is less than a certain threshold (preset to 0.4 times the average Euclidean distance, described as an example), it is considered that the two key points match. This is a well-known matching process. If the Any key point of the frame CT scan image is If there is no key point that is successfully matched on the frame CT scan image, then any key point will not be analyzed subsequently.

[0064] Get the CT scan image frame The position coordinates of the key points to the CT scan image frame The key point is The vector formed by the Euclidean distance and direction of the position coordinates of the key points that are successfully matched on the frame CT scan image is recorded as CT scan image frame The motion vector of each key point.

[0065] Thus, the motion vector of each key point in each frame of the CT scan image is obtained. In this embodiment, the motion vector of the key point in the last frame of the CT scan image is not analyzed.

[0066] The motion vectors of key points in all frames of CT scan images are embedded into a statistical iterative reconstruction algorithm (SIR) as prior knowledge to obtain an updated statistical iterative reconstruction algorithm.

[0067] Each frame of the CT scan image is input into the updated statistical iterative reconstruction algorithm, and a reconstructed CT scan image is output.

[0068] It should be noted that the statistical iterative reconstruction algorithm is an image reconstruction method based on a statistical model, which gradually approaches the real image through iterative optimization. In each iteration, the algorithm will use prior knowledge and observation data to update the image estimate until a certain stopping condition is met. This is a well-known technology and the specific method will not be introduced here. Embedding the motion vector of the key point in each frame of the image as prior knowledge into the statistical iterative reconstruction algorithm (SIR) is a well-known and feasible method. This method can use the motion information of the key point to improve the reconstruction quality of the video, that is, by modifying the cost function of the algorithm so that it takes into account the constraints of the motion vector, and thus using this prior knowledge in the reconstruction process, the static projection described by the system matrix of the traditional SIR is corrected to the motion-aware system matrix. Using the motion information of the key points can help the algorithm better understand the dynamic changes in the content of the scanned image, thereby improving the quality of the reconstructed image. Therefore, in this embodiment, by obtaining the accurate motion vector of the key point, the enhanced effect of the reconstructed CT scan image is guaranteed. The reconstructed CT image after removing the motion artifacts, such as Figure 4 shown.

[0069] The present invention also provides a system for improving CT image quality with micro-radiation dose, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned method for improving CT image quality with micro-radiation dose.

[0070] So far, the present invention is completed.

[0071] In summary, in an embodiment of the present invention, a plurality of consecutive CT scan images of the same part of a person being examined are acquired, a plurality of candidate matching points are screened from the feature points in each CT scan image frame, and a plurality of key points are then screened from the candidate matching points in each CT scan image frame. Based on the key point matching in adjacent CT scan image frames, the motion vectors of the key points in each CT scan image frame are acquired. The motion vectors of the key points in all CT scan image frames are embedded as prior knowledge into a statistical iterative reconstruction algorithm, and an updated statistical iterative reconstruction algorithm is acquired. Each CT scan image frame is input into the updated statistical iterative reconstruction algorithm, and a reconstructed CT scan image is output. By acquiring the precise motion vectors of the key points in the CT scan image, the present invention improves the quality of CT image reconstruction for micro-radiation doses.

[0072] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for improving the quality of CT images with micro-radiation dose, characterized in that: The method comprises the following steps: Obtaining a number of consecutive frames of CT scan images of the same part of the person being tested; Taking any frame of CT scan image as the target image, several frames of images with the same frequency are selected from the CT scan images other than the target image; obtaining several feature points in each frame of CT scan image, and selecting several candidate matching points from the feature points in the target image; According to the matching conditions of the candidate matching points in the target image and the same-frequency image, a number of key points are selected from the candidate matching points in the target image; According to the matching results of key points in adjacent frames of CT scan images, the motion vectors of the key points in each frame of CT scan image are obtained; the motion vectors of the key points in all frames of CT scan image are embedded as prior knowledge into the statistical iterative reconstruction algorithm to obtain an updated statistical iterative reconstruction algorithm; each frame of CT scan image is input into the updated statistical iterative reconstruction algorithm, and a reconstructed CT scan image is output.

2. The method for improving CT image quality with micro-radiation dose according to claim 1, characterized in that: The specific steps of selecting a plurality of frames of same-frequency images from the CT scan images other than the target image are as follows: In any row of any frame of CT scan image, the grayscale values of all pixels are counted from left to right to form a grayscale value sequence; All frames of CT scan images except the target image are recorded as reference images; Combine the target image with the The average of the inversely proportional normalized values of the DTW distances of the grayscale value sequences of all the same rows of the frame reference image is recorded as the distance between the target image and the first The overall image similarity of the frame reference image; According to the similarity between the target image and each frame of the reference image, the same-frequency image of the target image is determined.

3. The method for improving CT image quality with micro-radiation dose according to claim 2, characterized in that: The specific steps of determining the same-frequency image of the target image according to the overall similarity between the target image and each frame of the reference image are as follows: The reference image whose overall image similarity is greater than or equal to the preset similarity threshold is recorded as the same-frequency image of the target image.

4. The method for improving CT image quality with micro-radiation dose according to claim 1, characterized in that: The specific steps of selecting a plurality of candidate matching points from the feature points in the target image are as follows: For any feature point in the target image , in Acquisition and feature points in the same frequency image The feature point with the closest Euclidean distance to the position coordinates is recorded as the feature point Matching point, if the feature point With feature points When the Euclidean distance of the position coordinates of the matching point is less than the preset distance threshold, the feature point Recorded as the matching point to be selected.

5. The method for improving CT image quality with micro-radiation dose according to claim 4, characterized in that: The specific steps of selecting a number of key points from the candidate matching points in the target image are as follows: In the target image, each candidate matching point is used as the initial cluster center. According to the Euclidean distance between pixels, all pixels are clustered only once to obtain the connected area corresponding to each candidate matching point. For any connected region , to connect the regions The center point to the connected area The Euclidean distance and direction of each adjacent connected area center point are the vectors, which are recorded as connected areas. The corresponding neighborhood vector of the candidate matching point; Obtain the neighborhood vector of each candidate matching point in each frame of the same-frequency image according to the method of obtaining the neighborhood vector of each candidate matching point in the target image; According to the target image The first of the candidate matching points The neighborhood vector and the The candidate matching points are in The cosine similarity of each neighborhood vector of the matching point in the same-frequency image of the frame is used to determine the first The first of the candidate matching points The maximum similarity of neighborhood vectors; According to the maximum similarity, the first Filter out matching vectors from the neighborhood vectors of the candidate matching points; The target image The mean of the maximum similarity of all matching vectors of the candidate matching points is recorded as The candidate matching points are in Structural matching in frame-same frequency images; According to the structural matching degree of each candidate matching point in the target image in each frame of the same-frequency image, several key points are screened out.

6. The method for improving CT image quality with micro-radiation dose according to claim 5, characterized in that: The determining target image The first of the candidate matching points The maximum similarity of neighborhood vectors includes the following specific steps: The target image The first of the candidate matching points The neighborhood vector and the The candidate matching points are in The maximum value of the cosine similarity of all neighborhood vectors of the matching point in the same-frequency image is recorded as the first The first of the candidate matching points The maximum similarity of the neighborhood vectors.

7. The method for improving CT image quality with micro-radiation dose according to claim 5, characterized in that: According to the maximum similarity, the first The matching vectors are screened out from the neighborhood vectors of the candidate matching points. The specific steps are as follows: For the target image Among all the neighborhood vectors of the candidate matching points, the neighborhood vector with the maximum similarity greater than or equal to the preset vector threshold is recorded as the matching vector.

8. The method for improving CT image quality with micro-radiation dose according to claim 5, characterized in that: The specific steps of selecting a plurality of key points based on the structural matching degree of each candidate matching point in the target image in each frame of the same-frequency image are as follows: The product of the preset quantity coefficient and the number of images with the same frequency is recorded as the quantity threshold; For the target image The structural matching degree of the candidate matching points in all frames of the same-frequency images is calculated, and the same-frequency images with a structural matching degree greater than or equal to the preset matching threshold are recorded as matching same-frequency images. When the number of matching same-frequency images is greater than the number threshold, the first image in the target image is recorded as the matching same-frequency image. The matching points to be selected are recorded as key points.

9. The method for improving CT image quality with micro-radiation dose according to claim 1, characterized in that: The specific steps of obtaining the motion vector of the key point in each frame of the CT scan image are as follows: By brute force matching, get the Each key point of the frame CT scan image is The key points of successful matching on the frame CT scan image; Get the CT scan image frame The position coordinates of the key points to the CT scan image frame The key point is The vector formed by the Euclidean distance and direction of the position coordinates of the key points that are successfully matched on the frame CT scan image is recorded as CT scan image frame The motion vector of each key point.

10. A system for improving CT image quality by micro-radiation dose, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the steps of the method for improving the quality of CT images with micro-radiation doses as described in any one of claims 1 to 9 are implemented.

Images