CT image hybrid ring artifact removal method and device, terminal and medium

Abstract

The invention discloses a CT image hybrid ring artifact removal method and device, a terminal and a medium, and the method comprises the steps: extracting a structure image of an initial CT image after polar coordinate transformation, and determining a texture image according to the initial CT image and the structure image; classifying line segments in the texture image by adopting a trisection method, and determining a classification result; and correcting the initial CT image according to a classification result, and determining an artifact correction image. According to the method, the texture image is separated from the CT image, the line segments in the texture image are classified by adopting the trisection method, and the mixed / complex ring artifacts in the CT image are corrected based on the classification result instead of only processing the time-dependent ring artifacts, so that the processing accuracy of the CT image is improved. Therefore, the problems that when time-dependent ring artifacts are processed in the prior art, artifact correction is carried out only through the overall linear attribute of the ring artifacts, the artifact correction effect is poor easily, and new artifacts are possibly introduced can be effectively solved.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and particularly to a method, device, terminal and medium for removing mixed circular artifacts in CT images. Background Art

[0002] X-ray computed tomography (CT) imaging technology is widely used in fields such as disease diagnosis and industrial inspection. During the imaging process, CT images are easily affected by circular artifacts with different radii centered on the reconstruction center, thereby affecting the effects of disease diagnosis and industrial detection. The reasons for circular artifacts are more complex than expected, and the main reasons can be summarized as the intensity dependence of detector response and the time dependence of CT hardware systems, etc. Therefore, circular artifacts in CT images are divided into intensity-dependent circular artifacts, time-dependent circular artifacts, and their complex mixed forms. Effective correction of circular artifacts has a great impact on achieving high-quality CT imaging technology.

[0003] In recent years, circular artifact correction technologies based on image processing have been widely applied. Such technologies have no restrictions and requirements on the hardware operation level and can produce good artifact correction effects. According to the different data to be processed, CT image circular artifact correction technologies based on image processing can be divided into two categories: preprocessing and postprocessing technologies. Among them, the preprocessing method refers to processing and correcting abnormal data in projection data before reconstructing CT images using the projection data, while the postprocessing method refers to directly correcting artifacts in the reconstructed CT images. However, in the above-mentioned circular artifact correction technologies based on image processing, when dealing with time-dependent circular artifacts, only the overall linear attributes of circular artifacts are used to design the artifact correction algorithm, which easily leads to poor artifact correction effects and may introduce new artifacts. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method, device, terminal and medium for removing mixed circular artifacts in CT images aiming at the above-mentioned defects of the prior art, aiming to solve the problem that when dealing with time-dependent circular artifacts in the prior art, only the overall linear attributes of circular artifacts are used for artifact correction, which easily leads to poor correction effects and the introduction of new artifacts.

[0005] The technical solution adopted by the present invention to solve the problem is as follows:

[0006] In a first aspect, an embodiment of the present invention provides a method for removing mixed circular artifacts in CT images, where the method includes:

[0007] Extract a structure image based on the initial CT image under polar coordinate transformation, and determine a texture image according to the initial CT image and the structure image;

[0008] Classify the straight lines in the texture image using the trichotomy method to determine the classification result;

[0009] Correct the initial CT image according to the classification result to determine the artifact-corrected image.

[0010] In an implementation method, extract a structure image based on the initial CT image under polar coordinate transformation, including:

[0011] Convert the initial CT image to the polar coordinate system to determine the polar coordinate system image;

[0012] Extract the structure image from the polar coordinate system image based on a structure extraction model, where the structure extraction model consists of a fidelity term and the total window variation of pixel points.

[0013] In an implementation method, classify the straight lines in the texture image using the trichotomy method to determine the classification result, including:

[0014] Perform an initial classification of the texture image based on a straight line segment classification model to determine each straight line segment;

[0015] Iteratively optimize and classify each straight line segment using the trichotomy method to determine the classification result.

[0016] In an implementation method, the straight line segment classification model includes:

[0017]

[0018] Among them, represents the texture image under polar coordinate transformation, represents the value at the polar coordinate (r0, p), r0 represents the polar radius, p represents the polar angle, and K represents the number of straight line segment classifications, represents the mean value of each straight line segment, represents the demarcation point of each straight line segment, represents the k-th straight line segment.

[0019] In an implementation method, iteratively optimize and classify each straight line segment using the trichotomy method to determine the classification result, including:

[0020] Use the quartiles to divide each straight line segment into three sub-straight line segments;

[0021] When each sub-straight line segment is on the k-th class of straight line segments, determine and update the class of each sub-straight line segment according to the values of each sub-straight line segment, the mean value of the (k - 1)-th class of straight line segments, and the mean value of the (k + 1)-th class of straight line segments by comparing the numerical magnitudes.

[0022] Update the mean values and demarcation points corresponding to each type of straight line segment according to the types of the sub-straight line segments, and determine the classification result.

[0023] In one implementation method, the method further includes:

[0024] When the number of times of classifying each of the straight line segments by the trichotomy method is less than or equal to a preset number of times, use the sub-straight line segment as the straight line segment, and perform again the operation of classifying each of the straight line segments by the trichotomy method to determine the classification result.

[0025] In one implementation method, the classification result includes each of the straight line segments and the mean values and demarcation points respectively corresponding to each of the straight line segments. Correcting the initial CT image according to the classification result to determine an artifact-corrected image includes:

[0026] Correct the straight line segment based on the mean value and demarcation point corresponding to the straight line segment in the initial CT image to determine the artifact-corrected image.

[0027] In a second aspect, an embodiment of the present invention further provides a CT image hybrid circular artifact removal device. Among them, the CT image hybrid circular artifact removal device includes:

[0028] A texture image determination module, configured to extract a structure image based on an initial CT image under polar coordinate transformation, and determine a texture image according to the initial CT image and the structure image;

[0029] A classification result determination module, configured to classify a straight line in the texture image by the trichotomy method to determine a classification result;

[0030] An artifact-corrected image determination module, configured to correct the initial CT image according to the classification result to determine an artifact-corrected image.

[0031] In a third aspect, an embodiment of the present invention further provides a terminal. The terminal includes a memory and more than one processor; the memory stores more than one program; the program includes instructions for executing the CT image hybrid circular artifact removal method as described in any one of the above; the processor is configured to execute the program.

[0032] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which multiple instructions are stored. Among them, the instructions are suitable for being loaded and executed by a processor to implement the CT image hybrid circular artifact removal method as described in any one of the above.

[0033] Beneficial effects of the present invention: The embodiment of the present invention extracts a structural image based on an initial CT image under polar coordinate transformation, determines a texture image based on the initial CT image and the structural image, and realizes the conversion between ring artifacts and linear artifacts; classifies the straight lines in the texture image using the method of three to determine the classification result; corrects the initial CT image based on the classification result to determine the artifact correction image. Since the present invention separates the texture image from the CT image, classifies the straight lines in the texture image using the method of three, and corrects the mixed ring artifacts in the CT image based on the classification result, rather than just processing the time-dependent ring artifacts. Therefore, it can effectively solve the problems of the prior art that when processing time-dependent ring artifacts, only the overall linear properties of the ring artifacts are used to correct the artifacts, resulting in poor correction effect and the introduction of new artifacts. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 It is a flow chart of a method for removing mixed ring artifacts in CT images provided by an embodiment of the present invention.

[0036] Figure 2 Schematic diagram of ring artifacts in different coordinate systems provided by an embodiment of the present invention.

[0037] Figure 3 It is a schematic diagram of the ring artifact correction process provided by an embodiment of the present invention.

[0038] Figure 4 It is a schematic diagram of the internal modules of the device for removing mixed ring artifacts from CT images provided by an embodiment of the present invention.

[0039] Figure 5 It is a principle block diagram of a terminal provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The present invention discloses a method, device, terminal and medium for removing mixed ring artifacts from CT images. In order to make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0041] Those skilled in the art can understand that, unless specifically stated, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of the present invention means the presence of the stated features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0042] Those skilled in the art can understand that, unless otherwise defined, all terms used herein (including technical terms and scientific terms) have the same meaning as the general understanding of those of ordinary skill in the art to which the present invention pertains. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with an idealized or overly formal meaning unless specifically defined as here.

[0043] Ring artifact correction techniques based on image processing have been widely applied. Such techniques have no restrictions and requirements on the hardware operation level and can produce good artifact correction effects. According to the different data to be processed, the CT image ring artifact correction techniques based on image processing can be divided into two categories, such as preprocessing and postprocessing techniques. Among them, the preprocessing method refers to processing and correcting the abnormal data in the projection data before using the projection data for CT image reconstruction, while the postprocessing method refers to directly performing artifact correction on the reconstructed CT image. However, in the above-mentioned ring artifact correction techniques based on image processing, when dealing with time-dependent ring artifacts, only the overall linear attributes of the ring artifacts are used for artifact correction, which easily leads to problems such as poor correction effects and the introduction of new artifacts.

[0044] In view of the above defects of the prior art, the present invention provides a method for removing mixed circular artifacts in CT images. The method extracts a structure image from an initial CT image in the polar coordinate system, determines a texture image based on the initial CT image and the structure image, and realizes the conversion between circular artifacts and linear artifacts; uses the trichotomy method to classify the straight lines in the texture image to determine the classification result; corrects the initial CT image according to the classification result to determine an artifact-corrected image. Since the present invention separates the texture image from the CT image, uses the trichotomy method to classify the straight lines in the texture image, and corrects the mixed circular artifacts in the CT image based on the classification result, rather than only processing time-dependent circular artifacts. Therefore, it can effectively solve the problem that in the prior art, when dealing with time-dependent circular artifacts, only the overall linear attribute of the circular artifacts is used for artifact correction, which easily leads to poor correction effects and the introduction of new artifacts.

[0045] Exemplary method:

[0046] As Figure 1 shown, the method includes:

[0047] Step S100, extract a structure image from an initial CT image under polar coordinate transformation, and determine a texture image based on the initial CT image and the structure image.

[0048] For the initial CT image I under polar coordinate transformation p it can be decomposed into a structure image and a texture image That is wherein, the structure image contains the basic framework and layout of the image I p such as the geometric shape of an object and larger contours, etc. The texture image describes the fine structure of the local area of the image, such as cracks in the local area. The structure image has better smoothness in flat areas, while the texture image can provide detailed information about the area. Straight-line artifacts are an important part of the texture image. In the process of dealing with circular artifacts, the structure image can be extracted from the initial CT image under polar coordinate transformation, and then the texture image can be separated from the initial CT image based on the structure image, so as to correct the artifacts in the initial CT image by correcting the artifacts in the texture image. The texture image containing straight-line artifacts is decomposed through image separation technology, so as to effectively extract and locate abnormal straight-line values. The image separation technology utilizes the rapid mutation attribute of straight-line artifacts to accurately obtain the target values.

[0049] For the initial CT image I p , the structure image is obtained by eliminating the structure image p from the initial CT image I under polar coordinate transformation To obtain a texture image containing linear artifacts It can be expressed as:

[0050]

[0051] In one implementation, a structure image is extracted based on an initial CT image, including:

[0052] Step S101: Convert the initial CT image to the polar coordinate system to determine the polar coordinate system image;

[0053] Step S102: Extract the structure image from the polar coordinate system image based on a structure extraction model, where the structure extraction model consists of a fidelity term and the total window variation of pixel points.

[0054] Ring artifacts in a CT image appear as concentric circles centered on the reconstruction center, as Figure 2 shown. By using the geometric properties of a circle and the conversion between the Cartesian coordinate system and the polar coordinate system, ring artifacts can be converted into simpler linear artifacts. In the polar coordinate system, the initial CT image (polar coordinate system image) containing ring artifacts can be expressed as:

[0055] I p ={I p (r,θ)}, r ∈ {1, …, N r}}, θ ∈ {1, …, N θ};

[0056] where N r and N θ respectively represent the ranges of the polar radius and the polar angle. After coordinate transformation, ring artifacts appear as linear artifacts along the polar angle direction. Specifically, when the data of the polar radius r0 contains artifacts, it can be expressed as:

[0057]

[0058] where represents the ideal data, and δ(r0, θ) represents the linear artifact.

[0059] Through effective image smoothing and boundary preservation, the structure image can be extracted from the polar coordinate system image. Specifically, the structure extraction model can be used to extract the structure image from the polar coordinate image, where the structure extraction model has the following form:

[0060]

[0061] where the first term is the fidelity term, which avoids the structure image from being different from the input image I pThere are significant differences. The second term is the total window change, which is defined by the gradient change of the image and is used to measure the smoothness of the structural image. S is a potentially feasible structural image, and the optimal structural image is α is a parameter, ε represents a small positive number to prevent the denominator from being zero, D x , D y and L x , L y are defined by different change amounts, that is:

[0062]

[0063] Among them, represents the gradient of the image S in the y direction, N(i) represents the rectangular window area centered on i, and g i,j represents the Gaussian kernel function with i as the reference core. When the distance between j and i is large, g i,j has a small value. It can be seen that D x , D y and L x , L y can describe the total change amount of different pixel point windows (N(i)). The total window change amount has obvious numerical differences in the flat area and the boundary area of the image. D x , D y and L x , L y calculate the change amount in a slightly different way. L x , L y can better distinguish the boundary and background areas. Through the reasonable operation of these two types of change amounts, the contour change information of the structural image can be measured more effectively. Using the matrix attribute of the image, this non-convex optimization problem is transformed into a linear equation solving problem, and the structural image of the original artifact-damaged image I p can be obtained The structural image smooths the fine structure of the original image and only retains the structure of the larger areas in the image. Therefore, the subsequent texture image obtained based on this structural image contains rich line artifact information.

[0064] Step S200: Classify the lines in the texture image using the trichotomy method to determine the classification result.

[0065] According to the different causes of the circular artifacts, the line artifact δ(r0,θ) can generally be divided into two categories: intensity-dependent type and time-dependent type, etc. As Figure 3 shown, in the polar coordinate system, the intensity-dependent circular artifact is manifested as a line with approximately equal values along the polar angle direction, that is, δ(r0,θ1)≈δ(r0,θ2), θ1,θ2∈{1,…,N θ}. In contrast, the characteristics of time-dependent circular artifacts are more complex and usually appear as straight lines with similar numerical values along the polar angle direction, that is, the numerical values of the artifact δ(r0,θ) have piecewise similarity. There is an inclusion relationship between intensity-dependent linear artifacts and time-dependent linear artifacts. Considering that both piecewise continuous straight lines and overall continuous straight lines can be regarded as combinations of several straight lines. In the polar coordinate system, by using the discontinuous feature of the numerical values of stripe / linear artifacts, classifying the abnormal data into straight line segments and performing numerical correction can effectively remove time-dependent circular artifacts. There is an inclusion relationship between time-dependent circular artifacts and intensity-dependent circular artifacts. Therefore, this method is also applicable to the correction of intensity-dependent circular artifacts. In this embodiment, by using the texture image, the numerical discontinuity of the linear artifacts, and the consistency of the spatial positions, the abnormal linear artifacts in the texture image are classified based on the trichotomy method, so as to simultaneously process intensity-dependent circular artifacts and time-dependent circular artifacts and effectively remove complex mixed circular artifacts.

[0066] In one implementation, the trichotomy method is used to classify the straight lines in the texture image and determine the classification result, including:

[0067] Step S201: Initialize the classification of the texture image based on the straight line segment classification model to determine each straight line segment;

[0068] Step S202: Use the trichotomy method to perform iterative optimization and classification on each straight line segment to determine the classification result.

[0069] Specifically, first, the texture image is initialized and classified through the initialized straight line segment classification model to obtain each straight line segment; then, the trichotomy method is used to perform more detailed classification on each straight line segment again to obtain a more accurate classification result, avoiding the problem of spatial discontinuity in the classified straight line segments, that is, there may be data of other types of straight line segments in the k-th straight line segment .

[0070] In one implementation, the straight line segment classification model includes:

[0071]

[0072] Among them, represents the texture image under polar coordinate transformation, represents the numerical value at the polar coordinate (r0,p), r0 represents the polar radius, p represents the polar angle, K represents the number of classifications of the straight line segments, that is, the number of classifications of the linear artifacts, represents the mean value of each straight line segment, is an array of size K, represents the demarcation points of each straight line segment, is an array of size K + 1, represents the kth straight line segment, and Traverse the polar radius r∈N r This can realize the classification of all straight line artifacts in the texture image and obtain the mean value M of each classified straight line segment r and the starting point position P r And other information, so as to provide data preparation for the subsequent correction of abnormal straight line values.

[0073] In one implementation, the three-part method is used to iteratively optimize and classify each of the straight line segments to determine the classification result, including:

[0074] Dividing each of the straight line segments into three sub-segments using the tertile;

[0075] When each of the sub-segments is located on the k-th type of straight line segment, the category of each of the sub-segments is determined and updated by comparing the values of the sub-segments, the average of the k-1-th type of straight line segments, and the average of the k+1-th type of straight line segments;

[0076] According to the categories of the sub-straight line segments, the mean values and the dividing points corresponding to the various types of straight line segments are updated to determine the classification result.

[0077] In order to ensure the spatial continuity of the straight line segments during the classification process, each straight line segment is divided into three sub-segments, and the values of the three sub-segments are used to determine whether the two marginal sub-segments meet the category conversion conditions, that is, whether the first sub-segment should be classified as a straight line of the previous category, and whether the third sub-segment should be classified as a straight line of the latter category. This three-part classification method of straight line segments ensures the spatial continuity of the straight line segments while utilizing the numerical discontinuity of the straight line. Furthermore, each straight line segment can be divided into three sub-segments in an even manner.

[0078] Specifically, based on data As an example, the above steps are as follows:

[0079] Divide the space equally and obtain Initialize the classified data, including K types of straight line segments, and obtain the mean of each type of straight line segment data and the dividing point

[0080] For the kth straight line segment Segment it into three parts and record the first segmentation point as and the second split point is denoted as judge and Specifically:

[0081] when When it is divided into the (k - 1)-th type of straight line segments, otherwise it remains unchanged;

[0082] When it is divided into the (k + 1)-th type of straight line segments, otherwise it remains unchanged;

[0083] According to the classification results of each sub - straight line segment, update the mean value and the demarcation point of each type of straight line segment, so as to determine the classification results of each straight line segment by the trichotomy method for this time.

[0084] In one implementation manner, the method further includes:

[0085] When the number of times of classifying each of the straight line segments by the trichotomy method is less than or equal to a preset number of times, use the sub - straight line segments as the straight line segments, and perform again the operation of classifying each of the straight line segments by the trichotomy method to determine the classification results.

[0086] To improve the accuracy of the classification results and avoid excessive consumption of computing resources, classify each straight line segment by the trichotomy method multiple times, and set a preset number of times to limit the number of times of performing the operation of classifying each straight line segment by the trichotomy method. During the execution, obtain the number of times of currently executing the operation of "classifying each of the straight line segments by the trichotomy method to determine the classification results". When this number is less than or equal to the preset number of times, use the sub - straight line segments obtained by segmentation in the previous operation as the straight line segments, and perform again the step of "classifying each of the straight line segments by the trichotomy method to determine the classification results" for each straight line segment until the number of executions is greater than the preset number of times, then stop executing the operation of "classifying each of the straight line segments by the trichotomy method to determine the classification results", and use the current classification results as the final classification results.

[0087] When performing the operation of "classifying each of the straight line segments by the trichotomy method to determine the classification results", its steps include: dividing each of the straight line segments into three sub - straight line segments; when each of the sub - straight line segments is on the k - th type of straight line segments, according to the values of each of the sub - straight line segments, the mean value of the (k - 1)-th type of straight line segments, and the mean value of the (k + 1)-th type of straight line segments, use the comparison of numerical magnitudes to determine and update the categories of each of the sub - straight line segments; update the mean values and demarcation points corresponding to each type of straight line segments according to the categories of each of the sub - straight line segments to determine the classification results.

[0088] In one implementation manner, it is also possible to judge whether to perform again the operation of "classifying each of the straight line segments by the trichotomy method to determine the classification results" by setting an exit condition. Specifically as follows: set an exit threshold A, when or When it is, perform the operation of "classifying each of the straight line segments using the trichotomy method to determine the classification result"; otherwise, exit the execution.

[0089] Step S300: Correct the initial CT image according to the classification result to determine the artifact-corrected image.

[0090] After the classification of the linear artifact segments is completed, the correction of the annular artifacts can be achieved based on the classification result and the numerical continuity of the straight lines. The internal values of each classified straight line segment are approximately equal. Therefore, the mean value of the straight line segment can be used to accurately correct the artifacts of each straight line segment.

[0091] In one implementation, the classification result includes each of the straight line segments and the mean value and the demarcation point respectively corresponding to each of the straight line segments. Correcting the initial CT image according to the classification result to determine the artifact-corrected image includes:

[0092] Correcting the straight line segment based on the mean value and the demarcation point corresponding to the straight line segment in the initial CT image to determine the artifact-corrected image.

[0093] Simply put, the values on the same straight line segment after classification are approximately equal. Therefore, the range of the straight line segment can be defined by the demarcation point, and the values of the straight line segment can be corrected by the mean value of the straight line segment. After all the classified straight line segments in the initial CT image are corrected, the artifact-corrected image can be obtained. Specifically, denote the classification result of the linear artifact as F(K,M r ,P r ), r ∈ N r , then the corrected polar coordinate CT image (artifact-corrected image) is:

[0094]

[0095] s.t. θ ∈ [P r (t - 1), P r (t)];

[0096] where θ ∈ [P r (t - 1), P r (t)] means that θ is an element of the t-th type of straight line segment.

[0097] Based on the above embodiments, the present invention also provides a CT image hybrid annular artifact removal device, as Figure 4 shown, the device includes:

[0098] A texture image determination module 01, configured to extract a structure image based on the initial CT image in polar coordinates, and determine a texture image according to the initial CT image and the structure image;

[0099] The classification result determination module 02 is configured to classify the straight lines in the texture image by using the trichotomy method to determine the classification result;

[0100] The corrected image determination module 03 is configured to correct the initial CT image according to the classification result to determine the artifact-corrected image.

[0101] Based on the above embodiments, the present invention further provides a terminal, and its principle block diagram can be as Figure 5 shown. The terminal includes a processor, a memory, a network interface, and a display screen connected through a system bus. Among them, the processor of the terminal is used to provide computing and control capabilities. The memory of the terminal includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the terminal is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it is used to implement the method for removing mixed circular artifacts in CT images. The display screen of the terminal can be a liquid crystal display screen or an electronic ink display screen.

[0102] Those skilled in the art can understand that Figure 5 the principle block diagram shown in

[0103] merely shows the block diagram of some structures related to the solution of the present invention, and does not constitute a limitation on the terminal to which the solution of the present invention is applied. The specific terminal may include more or fewer components than those shown in the figure, or combine some components, or have a different component layout.

[0104] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided by the present invention can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0105] In summary, the present invention discloses a method, device, terminal, and medium for removing mixed circular artifacts in CT images. The method extracts a structure image based on an initial CT image in a polar coordinate system, determines a texture image according to the initial CT image and the structure image, and realizes the conversion of circular artifacts and linear artifacts; uses the trichotomy method to classify the lines in the texture image to determine the classification result; corrects the initial CT image according to the classification result to determine an artifact-corrected image. Since the present invention separates the texture image from the CT image, uses the trichotomy method to classify the line segments in the texture image, and corrects the mixed circular artifacts in the CT image based on the classification result, rather than only processing time-dependent circular artifacts. Therefore, it can effectively solve the problems in the prior art that when processing time-dependent circular artifacts, only the overall linear attribute of the circular artifacts is used for artifact correction, which easily leads to poor correction effects and the introduction of new artifacts.

[0106] It should be understood that the application of the present invention is not limited to the above examples. For those of ordinary skill in the art, improvements or transformations can be made according to the above description. All such improvements and transformations should fall within the protection scope of the appended claims of the present invention.

Claims

1. A method for removing mixed circular artifacts in CT images, characterized in that, The method includes: extracting a structure image based on an initial CT image under polar coordinate transformation, and determining a texture image according to the initial CT image and the structure image; classifying the straight lines in the texture image by using the trichotomy method to determine a classification result; correcting the initial CT image according to the classification result to determine an artifact-corrected image.

2. The CT image hybrid circular artifact removal method according to claim 1, wherein Extracting a structure image based on an initial CT image under polar coordinate transformation includes: converting the initial CT image to a polar coordinate system to determine a polar coordinate system image, and realizing the conversion of circular artifacts and linear artifacts; extracting the structure image from the polar coordinate system image based on a structure extraction model, where the structure extraction model is composed of a fidelity term and the total window variation of pixel points.

3. The CT image hybrid circular artifact removal method according to claim 1, wherein, Classifying the straight lines in the texture image by using the trichotomy method to determine a classification result includes: performing an initial classification on the texture image based on a straight-line segment classification model to determine each straight-line segment; performing iterative optimization and classification on each straight-line segment by using the trichotomy method to determine a classification result.

4. The CT image hybrid circular artifact removal method according to claim 3, wherein, The straight-line segment classification model includes: Among them, represents the texture image under polar coordinate transformation, represents the value at the polar coordinate (r0, p), where r0 represents the polar radius, p represents the polar angle, and K represents the number of classifications of the straight line segments, represents the mean value of each of the straight line segments, represents the demarcation point of each of the straight line segments, represents the k-th straight line segment.

5. The CT image hybrid circular artifact removal method according to claim 3, characterized in that, Classifying each straight-line segment by using the trichotomy method to determine a classification result includes: dividing each straight-line segment into three sub-straight-line segments by using the three-quantile method; when each sub-straight-line segment is located on the k-th type of straight-line segment, determining and updating the category of each sub-straight-line segment according to the values of each sub-straight-line segment, the mean value of the (k - 1)-th type of straight-line segment, and the mean value of the (k + 1)-th type of straight-line segment by using the comparison of numerical magnitudes; updating the mean value and the demarcation point corresponding to each type of straight-line segment according to the category of each sub-straight-line segment to determine the classification result.

6. The CT image hybrid circular artifact removal method according to claim 5, wherein The method further includes: when the number of times of classifying each straight-line segment by using the trichotomy method is less than or equal to a preset number of times, taking the sub-straight-line segment as the straight-line segment, and performing again the operation of classifying each straight-line segment by using the trichotomy method to determine the classification result.

7. The CT image hybrid circular artifact removal method according to claim 5, wherein The classification result includes each straight-line segment and the mean value and the demarcation point respectively corresponding to each straight-line segment. Correcting the initial CT image according to the classification result to determine an artifact-corrected image includes: correcting the straight-line segment based on the mean value and the demarcation point corresponding to the straight-line segment in the initial CT image to determine the artifact-corrected image.

8. A CT image hybrid circular artifact removal device, characterized in that The device includes: a texture image determination module, configured to extract a structure image based on an initial CT image under polar coordinate transformation, and determine a texture image according to the initial CT image and the structure image; a classification result determination module, configured to classify the straight lines in the texture image by using the trichotomy method to determine a classification result; a corrected image determination module, configured to correct the initial CT image according to the classification result to determine an artifact-corrected image.

9. A terminal, characterized in that, The terminal includes a memory and more than one processor; the memory stores more than one program; the program includes instructions for executing the CT image hybrid circular artifact removal method according to any one of claims 1 - 7; the processor is configured to execute the program.

10. A computer-readable storage medium storing multiple instructions, characterized in that, The instruction is loaded and executed by a processor to implement the steps of the CT image hybrid circular artifact removal method according to any one of claims 1-7 above.

Images