Image processing device and x-ray CT device

Abstract

To perform motion compensation.SOLUTION: An image processing method according to an embodiment includes the steps of: receiving projection data acquired from an imaging object by an X-ray CT (computed tomography) device; until a predefined termination criterion is met, iteratively reconstructing, based on estimated cardiac motion, the received projection data to generate a motion-compensated image of the imaging object; determining a vessel region of interest (ROI) within the generated motion-compensated image; updating the estimated cardiac motion, based on an optimization cost function associated with the determined vessel ROI; and outputting, as a final reconstructed image of the imaging object, the generated motion-compensated image.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The embodiments disclosed in this specification and the drawings relate to an image processing method and an X-ray CT apparatus.

Background Art

[0002] Cardiac CT (Computed Tomography) is one of the most difficult fields in medical imaging because the heart is constantly moving and has both regular and irregular movement patterns. Therefore, the latest imaging technology is required to capture clear cardiac images while the heart and blood vessels are moving.

[0003] Typically, compensation methods for restoring image quality are applied to reduce the impact of artifacts caused by heart movement. Many of these methods have two main stages: (1) estimation of heart movement, and (2) incorporation of the estimated movement into the image reconstruction process to address motion artifacts. The performance of these motion compensation methods is mainly determined by the accuracy of the motion estimation stage.

[0004] It is desirable to enhance current cardiac motion compensation approaches.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to perform motion compensation. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problems. The problems corresponding to the respective effects of the respective configurations shown in the embodiments described later can also be regarded as other problems.

Means for Solving the Problems

[0007] The image processing method according to the embodiment receives projection data obtained from an imaging object by an X-ray CT (Computed Tomography) apparatus, and repeatedly reconstructs the projection data based on the estimated motion of the heart until a predetermined end criterion is satisfied, to generate a motion-compensated image of the imaging object, determines a region of interest (ROI) of blood vessels in the motion-compensated image, updates the motion of the heart based on an optimization cost function related to the determined blood vessel ROI, and outputs the generated motion-compensated image as the finally reconstructed image of the imaging object.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments of an image processing method and an X-ray CT apparatus will be described in detail with reference to the drawings.

[0010] The following disclosure presents many different embodiments, or examples, for implementing various features of the presented subject matter. Specific examples of components and arrangements are described below to simplify the present disclosure. These are, of course, merely examples and are not intended to be limiting.

[0011] For example, as described herein, the order of the description of the different steps is presented for clarity. Generally, these steps can be performed in any suitable order. Also, each of the different features, techniques, and configurations in this specification may be described in various places in the present disclosure, but each concept is intended to be executable independently of or in combination with each other. Accordingly, the present disclosure can be embodied and considered in many different ways.

[0012] The present disclosure relates to methods and apparatuses that can integrate motion estimation and motion compensation rather than performing them in two independent stages. By doing so, the accuracy of motion estimation can be improved during motion compensation, and the efficiency of motion compensation can be improved through the motion estimation process. The present disclosure also introduces a method for utilizing vascular segmentation in a combined estimation-compensation procedure. Furthermore, the present disclosure provides a detailed workflow for implementing this framework.

[0013] FIG. 1 shows a block diagram of an exemplary apparatus 100 for performing cardiac motion estimation and compensation in a CT imaging system (X-ray CT apparatus) according to an embodiment of the present disclosure. The apparatus 100 includes a projection data reception circuit 110, an image reconstruction circuit 120, a vascular region of interest (ROI) determination circuit 130, a cardiac motion update circuit 140, and a reconstructed image output circuit 150.

[0014] The projection data reception circuit receives unprocessed projection data obtained from an imaging target by the CT imaging system and transmits the data to the image reconstruction circuit 120.

[0015] Based on the estimated heart movement, the image reconstruction circuit 120 reconstructs the projection data to generate a motion-compensated image of the imaging object, and transmits the generated image to the vascular ROI determination circuit 130.

[0016] The vascular ROI determination circuit 130 determines the vascular ROI in the generated motion-compensated image, and transmits the determined vascular ROI to the heart motion update circuit 140.

[0017] Based on the optimization cost function related to the vascular ROI received from the vascular ROI determination circuit 130, the heart motion update circuit 140 updates the estimated heart motion, and transmits the updated heart motion to the image reconstruction circuit 120. Next, the updated heart motion is used in the next pass of image reconstruction.

[0018] The image reconstruction circuit 120, the vascular ROI determination circuit 130, and the heart motion update circuit 140 perform motion estimation and motion compensation in a repetitive manner until a predetermined end criterion is met. Based on the optimization cost function, a number of optimization methods including, but not limited to, the gradient descent method, the stochastic gradient descent method, the Adam method, and the Newton method can be used to repeatedly update the motion.

[0019] When the end criterion is met, the reconstructed image output circuit 150 outputs the motion-compensated image generated in the last iteration as the final image of the imaging object.

[0020] FIG. 2 shows a flowchart of an exemplary procedure 200 for performing heart motion estimation and compensation in a CT imaging system according to an embodiment of the present disclosure. As shown in FIG. 2, the procedure 200 starts at step S210 of receiving the unprocessed projection data of the imaging object.

[0021] At step S220, based on the estimated heart motion, the received projection data is reconstructed to generate a motion-compensated image of the imaging object.

[0022] In step S230, the vascular ROI is determined within the generated motion-compensated image.

[0023] In step S240, the estimated heart motion is updated based on an optimization cost function associated with the determined vascular ROI.

[0024] In step S250, it is determined whether a predetermined end criterion is satisfied. The operations in steps S220, S230, S240, and S250 are repeatedly performed until the end criterion is satisfied. If the determination result in step S250 is "yes", procedure 200 proceeds to step S260.

[0025] In step S260, the generated motion-compensated image is output as the final reconstructed image of the imaging target.

[0026] Hereinafter, different embodiments of the workflow for implementing the combined motion estimation-compensation procedure will be described in detail with reference to different embodiments.

[0027] FIG. 3 shows a block diagram of an exemplary apparatus 300 for performing heart motion estimation and compensation in a CT imaging system according to an embodiment of the present disclosure. The apparatus 300 includes a projection data reception circuit 310, an image reconstruction circuit 320, a vascular ROI determination circuit 330, a loss-based motion update circuit 340, and a reconstructed image output circuit 350.

[0028] As shown in FIG. 3, the image reconstruction circuit 320 includes a motion compensation (MC) -less image reconstruction circuit 322 and an MC image reconstruction circuit 324, and the vascular ROI determination circuit 330 includes a vascular segmentation circuit 332 and a vascular ROI identification circuit 334. The projection data reception circuit 310 and the reconstructed image output circuit 350 can be the same as the corresponding components of the apparatus 100 shown in FIG. 1, and therefore, additional descriptions are omitted to avoid redundancy.

[0029] The non-MC image reconstruction circuit 322 reconstructs the projection data received from the projection data receiving circuit 310 and generates an image of the imaging target without motion compensation. Next, the non-MC image reconstruction circuit 322 transmits the generated image to the blood vessel segmentation circuit 332.

[0030] The blood vessel segmentation circuit 332 performs blood vessel segmentation on the image received from the non-MC image reconstruction circuit 322 and creates a blood vessel region mask. The blood vessel region mask can cover the blood vessel tree structure and the range of motion of the blood vessel tree structure. For a detailed description of the blood vessel region mask, see FIGS. 4 to 5, which will be described later.

[0031] Based on the estimated motion of the heart, the MC image reconstruction circuit 324 reconstructs the projection data received from the projection data receiving circuit 310 to generate a motion-compensated image of the imaging target, and then transmits it to the blood vessel ROI identification circuit 334.

[0032] Based on the blood vessel region mask received from the blood vessel segmentation circuit 332, the blood vessel ROI identification circuit 334 identifies the blood vessel ROI in the motion-compensated image.

[0033] The loss-based motion update circuit 340 receives the blood vessel ROI identified by the blood vessel ROI identification circuit 334. Based on the loss function of the blood vessel ROI, the loss-based motion update circuit 340 updates the estimated motion of the heart. The updated motion of the heart is transmitted again to the MC image reconstruction circuit 324 for use in the next iteration.

[0034] As described above, the blood vessel region mask can cover the blood vessel tree structure and the range of motion of this structure. The creation of the blood vessel region mask can be divided into the following three steps.

[0035] The first step includes obtaining a cardiovascular skeleton composed of the centerlines of the blood vessels. For example, one approach is to apply a vascular filter to the image reconstructed without motion compensation so as to extract the cardiovascular skeleton. Various other blood vessel segmentation methods can be used, including but not limited to dynamic growth methods, AI-based deep learning, etc. FIG. 4 shows an exemplary cardiovascular skeleton extracted based on the image reconstructed without motion compensation according to one embodiment of the present disclosure.

[0036] Second, a vessel template T(x, y, z) can be derived based on the cardiovascular skeleton. The vessel template can represent a vessel tree structure without motion. In the absence of motion, the blood vessels take a tubular shape. Therefore, tubular regions can be generated for each branch of the cardiovascular skeleton, and the overall cardiovascular tree structure is formed as a whole. At the starting point of the branch, the tubular region may have a predetermined radius. Along the axial direction of the branch, the radius of the tubular region may gradually decrease from the starting point to the ending point. Each tubular region is constructed to mimic the corresponding blood vessel with the common Hounsfield Unit (HU) value of that blood vessel assigned to the region. FIG. 5 shows an exemplary cardiovascular tree created based on the extracted cardiovascular skeleton according to one embodiment of the present disclosure. This tubular blood vessel tree can be used as the vessel template T(x, y, z).

[0037] However, due to the movement of the blood vessels, the ideal tubular shape cannot be maintained. Therefore, the third step includes creating a vessel region mask M(x, y, z) based on the vessel template T(x, y, z). This mask M(x, y, z) encompasses both the vessel template and the range of its motion. That is, the mask not only covers only the tubular regions, but also mimics the blood vessels without ideal motion, and is large enough to cover the entire region contaminated or affected by the movement of the blood vessels.

[0038] FIG. 6 shows a flowchart of an exemplary procedure 600 for performing heart motion estimation and compensation in a CT imaging system according to an embodiment of the present disclosure.

[0039] In step S610, unprocessed projection data of the imaging object is received. In step S620, the received projection data is reconstructed to generate an image without motion compensation. In step S630, vessel segmentation is performed on the reconstructed image without motion compensation to create a vessel region mask M(x, y, z).

[0040] In step S640, based on the estimated heart motion, the projection data is reconstructed to generate a motion-compensated image of the imaging object. For example, to show the heart motion during acquisition of the projection data of the imaging object by the CT imaging system, a four-dimensional non-rigid motion V(x, y, z, t) can be estimated. During the first iteration, the estimated motion V(x, y, z, t) can be set to a predetermined value, for example, 0. By incorporating the estimated motion V(x, y, z, t) into the image reconstruction, a motion-compensated image of the imaging object can be obtained.

[0041] In step S650, based on the vessel region mask M(x, y, z) created in step S630, a vessel ROI is identified within the reconstructed motion-compensated image. In step S660, the heart motion is calculated to minimize the loss function of the vessel ROI, and then the estimated heart motion V(x, y, z, t) is updated using the calculated heart motion. To reduce the blur within the vessel ROI, the optimization loss function can be based on, for example, the entropy within the vessel ROI, the edge function of the vessel ROI, or a combination of both.

[0042] For example, the loss function may be formulated based on the gradient to explain the edge information within the vascular ROI. Alternatively or additionally, histogram-based entropy can be used to characterize the luminance of the vascular ROI. As the optimization progresses, the edges within the vascular ROI become sharper, and for example, the vascular ROI may exhibit an increase in luminance.

[0043] In step S670, it is determined whether a predetermined end criterion is satisfied. The operations from steps S640 to S670 are repeatedly executed until the predetermined end criterion is satisfied. For example, the end criterion may be that the loss function converges to a predetermined threshold or the number of repetitions reaches a predetermined value.

[0044] For example, the repetitive operation may end either when the loss change is within less than 0.1% or when the number of repetitions reaches 50.

[0045] If it is determined in step S670 that the end criterion is satisfied, procedure 600 proceeds to step S680. In step S680, the motion-compensated image generated in the last iteration is output as the final image of the imaging target.

[0046] FIG. 7 shows a block diagram of an exemplary apparatus 700 for performing heart motion estimation and compensation in a CT imaging system according to an embodiment of the present disclosure. FIG. 8 shows a flowchart of an exemplary procedure 800 for performing heart motion estimation and compensation in a CT imaging system according to an embodiment of the present disclosure. The only difference between the embodiments shown in FIGS. 7 to 8 from the embodiments shown in FIGS. 3 and 6 is that the vascular region mask is created based on the motion-compensated image instead of the non-motion-compensated image. This means that in the embodiments of FIGS. 7 to 8, the created vascular region mask can be continuously updated during the repetitive operation.

[0047] As another approach, the heart motion may be optimized based on other measurement criteria. For example, in the following embodiments, the similarity between the vascular ROI and the vascular template is used as the optimization measurement criterion.

[0048] FIG. 9 shows a block diagram of an exemplary apparatus 900 for performing heart motion estimation and compensation in a CT imaging system according to an embodiment of the present disclosure. The apparatus 900 includes a projection data receiving circuit 910, an image reconstruction circuit 920, a vascular ROI determination circuit 930, a similarity-based motion update circuit 940, and a reconstructed image output circuit 950.

[0049] Since the projection data receiving circuit 910, the image reconstruction circuit 920, and the reconstructed image output circuit 950 are the same as the projection data receiving circuit 310, the image reconstruction circuit 320, and the reconstructed image output circuit 350 of the apparatus 300 shown in FIG. 3, additional description is omitted.

[0050] The vascular ROI determination circuit 930 includes a vascular segmentation circuit 932 and a vascular ROI identification circuit 934. The vascular segmentation circuit 932 performs vascular segmentation on the generated motion-compensated-free image to create a vascular template and a vascular region mask. The vascular segmentation circuit 932 transmits the vascular template to the similarity-based motion update circuit 940 and transmits the vascular region mask to the vascular ROI identification circuit 934.

[0051] Similarly, in the embodiments of FIGS. 9 to 10, vascular segmentation may be performed on the motion-compensated image as in the embodiments shown in FIGS. 7 to 8, instead of the motion-compensated-free image.

[0052] Based on the received vascular region mask, the vascular ROI identification circuit 934 identifies the vascular ROI in the motion-compensated image generated by the MC-with image reconstruction circuit 924.

[0053] The similarity-based motion update circuit 940 receives the vascular ROI from the vascular ROI identification circuit 934 and receives the vascular template from the vascular segmentation circuit 932. The similarity-based motion update circuit 940 calculates the motion of the heart so as to maximize the similarity between the vascular ROI and the vascular template, and updates the estimated heart motion to the calculated heart motion. Next, the similarity-based motion update circuit 940 transmits the updated heart motion to the MC-with-image reconstruction circuit 924 for use in the next pass of image reconstruction.

[0054] Figure 10 shows a flowchart of an exemplary procedure 1000 for performing heart motion estimation and compensation in a CT imaging system according to an embodiment of the present disclosure. At step S1010, the unprocessed projection data of the imaging subject is received. At step S1020, the projection data is reconstructed to generate an image of the imaging subject without motion compensation. At step S1030, vascular segmentation is performed on the reconstructed image without motion compensation to create a vascular template and a vascular region mask.

[0055] The operations of steps S1040 to S1070 are repeatedly executed until a predetermined end criterion is satisfied. For example, the end criterion may be that the similarity converges to a predetermined threshold value or the number of repetitions reaches a predetermined value.

[0056] At step S1040, based on the estimated heart motion, the received projection data is reconstructed to generate a motion-compensated image. At step 1050, based on the vascular region mask, the vascular ROI is identified within the reconstructed motion-compensated image. At step S1060, the estimated heart motion is updated to the heart motion that maximizes the similarity between the vascular template and the vascular ROI.

[0057] There are various measurement criteria available for quantifying the similarity between a vascular template extracted from a motion-compensated image and a vascular ROI. Common measurement criteria include Euclidean distance, Pearson's correlation, and cross-correlation, among others. For example, the goal of the optimization process may be to maximize the cross-correlation, thereby potentially aligning the vascular template as closely as possible with the ROI region within the motion-compensated image. A cross-correlation value of 1.0 may imply an exact match, while a value of 0 implies a complete difference.

[0058] If it is determined in step S1070 that the end criterion is satisfied (i.e., the "yes" branch of S1070), procedure 1000 proceeds to step S1080. In step S1080, the motion-compensated image generated during the last iteration is output as the final image of the imaging subject.

[0059] Alternatively, the motion of the heart may be optimized using a cost function based on both the loss function of the vascular SOI and the similarity between the vascular template and the vascular SOI. For example, the loss function and the similarity can be incorporated as weighted terms within the cost function. The loss function may correspond to the data term for successive optimizations, while the similarity may potentially correspond to the regularization term to avoid overfitting. It should be noted that all of the cost functions described in this disclosure are exemplary and not limiting. One of ordinary skill in the art will understand that other forms of optimization measurement criteria are possible.

[0060] FIG. 11 shows a block diagram of an exemplary apparatus 1100 for performing heart motion estimation and compensation in a CT imaging system according to an embodiment of the present disclosure. The projection data receiving circuit 1110, the image reconstruction circuit 1120, and the reconstructed image output circuit 1150 of the apparatus 1100 may be the same as the projection data receiving circuit 310, the image reconstruction circuit 320, and the reconstructed image output circuit 350 of the apparatus 300 shown in FIG. 3, and the vascular segmentation circuit 1130 may be the same as the vascular segmentation circuit 932 of the apparatus 900 shown in FIG. 9.

[0061] The weighted cost-based motion update circuit 1140 receives a vascular ROI from the vascular ROI specifying circuit 1134 and a vascular template from the vascular segmentation circuit 1132. The weighted cost-based motion update circuit 1140 calculates the motion of the heart to optimize a cost function based on a loss function of the vascular ROI and a weighted term including the similarity between the vascular ROI and the vascular template, and updates the estimated heart motion to the calculated heart motion. Next, the weighted cost-based motion update circuit 1140 transmits the updated heart motion to the MC-with-image reconstruction circuit 1124 for use in the next pass of image reconstruction.

[0062] FIG. 12 shows a flowchart of an exemplary procedure 1200 for performing heart motion estimation and compensation in a CT imaging system according to an embodiment of the present disclosure.

[0063] In step S1210, unprocessed projection data of the imaging subject is received. In step S1220, the received projection data is reconstructed to generate a motion-compensation-free image of the imaging subject. In step S1230, vascular segmentation is performed on the motion-compensation-free reconstructed image to create a vascular template and a vascular region mask.

[0064] The operations from step S1240 to S1270 are repeatedly performed until a predetermined end criterion is satisfied. For example, the end criterion may be that the cost function converges to a predetermined threshold value or the number of repetitions reaches a predetermined value.

[0065] In step S1240, based on the estimated heart motion, the received projection data is reconstructed to generate a motion-compensated image. In step S1250, based on the vascular region mask, a vascular ROI is identified within the reconstructed motion-compensated image. In step S1260, the estimated heart motion is updated with the heart motion that optimizes a cost function based on the loss function of the vascular SOI and the similarity between the vascular SOI and the vascular template.

[0066] If it is determined in step S1270 that the end criterion is satisfied (i.e., the "yes" branch of S1270), procedure 1200 proceeds to step S1280. In step S1280, the motion-compensated image generated during the last repetition is output as the final image of the imaging object.

[0067] Again, in the embodiments of FIGS. 11 to 12, vascular segmentation may be performed on the motion-compensated image as in the embodiments shown in FIGS. 7 to 8, instead of the image without motion compensation.

[0068] Although the above embodiments are described in relation to cardiac CT imaging, those skilled in the art will understand that this approach is applicable to imaging other vascular structures without departing from the spirit and scope of the present disclosure.

[0069] FIG. 13 is a schematic block diagram of a CT apparatus or scanner according to an embodiment of the present disclosure. As shown in FIG. 13, which is depicted as seen from the side, the radiation imaging gantry 1350 further includes an X-ray tube 1351, an annular frame 1352, and a multi-row or two-dimensional array type X-ray detector 1353. The X-ray tube 1351 and the X-ray detector 1353 are placed on the annular frame 1352 in exactly opposite positions across the subject OBJ, and the annular frame 1352 is rotatably supported about a rotation axis RA. The rotating device 1357 rotates the annular frame 1352 at a high speed such as 0.4 seconds / rotation, and at the same time, the subject OBJ is moved along the axis RA in the direction behind or in front of the illustrated plane.

[0070] Embodiments of the X-ray CT apparatus according to the present disclosure will be described below with reference to the accompanying drawings. Note that the X-ray CT apparatus includes various types of apparatuses such as a rotation / rotation type apparatus in which both the X-ray tube and the X-ray detector rotate around the subject to be examined, and a fixed / rotation type apparatus in which a large number of detector elements are arranged in an annular or horizontal shape and only the X-ray tube rotates around the subject to be examined. The present disclosure is applicable to any type. Here, an example of the currently mainstream rotation / rotation type will be illustrated.

[0071] The multi-slice X-ray CT apparatus further includes a high voltage generator 1359 that generates a tube voltage applied to the X-ray tube 1351 through a slip ring 1358 so that the X-ray tube 1351 generates X-rays. The X-rays are radiated toward the subject OBJ, and the cross-sectional area thereof is represented by a circle. For example, the average X-ray energy during the first scan of the X-ray tube 1351 is less than the average X-ray energy during the second scan. Therefore, two or more scans corresponding to different X-ray energies can be acquired. The X-ray detector 1353 is arranged on the opposite side of the X-ray tube 1351 across the subject OBJ to detect the radiation X-rays that have propagated through the subject OBJ. The X-ray detector 1353 further includes individual detector elements or units.

[0072] The CT apparatus further includes other devices that process detection signals from the X-ray detector 1353. The data acquisition circuit or data acquisition system (DAS) 1354 converts the signal output from the X-ray detector 1353 for each channel into a voltage signal, amplifies the signal, and further converts the signal into a digital signal. The X-ray detector 1353 and the DAS 1354 are configured to process a predetermined total number of projections per rotation (TPPR).

[0073] The above data is transmitted through the non-contact data transmitter 1355 to the preprocessing device 1356 housed in the console outside the radiographic gantry 1350. The preprocessing device 1356 performs specific corrections such as sensitivity correction on the unprocessed data. The memory 1362 stores the resulting data, also called projection data, at a stage immediately before the reconstruction process. The memory 1362 is connected to the system controller 1360 through the data / control bus 1361 together with the reconstruction device 1364, the input device 1365, and the display 1366. The system controller 1360 controls the current regulator 1363 that limits the current to a level sufficient to drive the CT system.

[0074] The detector rotates and / or is fixed with respect to the patient, regardless of the generation of the CT scanner system. In one implementation, the above CT system may be an example in which a third-generation geometry system and a fourth-generation geometry system are combined. In the third-generation system, the X-ray tube 1351 and the X-ray detector 1353 are placed exactly opposite each other on the annular frame 1352 and rotate around the subject OBJ when the annular frame 1352 rotates about the rotation axis RA. In the fourth-generation geometry system, the detector is fixedly installed around the patient, and the X-ray tube rotates around the patient. In an alternative implementation, the radiographic gantry 1350 has a large number of detectors arranged on the annular frame 1352, which is supported by a C-arm and a stand.

[0075] The memory 1362 can store measurement values indicating the X-ray irradiation dose in the X-ray detector 1353. Further, the memory 1362 can store a dedicated program for performing CT image reconstruction, material decomposition, and motion estimation and compensation methods including the methods described in this specification.

[0076] The reconstruction device 1364 can execute the above reference methods described in this specification. Further, the reconstruction device 1364 can execute pre-reconstruction image processing such as volume rendering processing and image difference processing as necessary.

[0077] The pre-reconstruction processing of the projection data performed by the preprocessing device 1356 may include, for example, correcting detector calibration, detector non-linearity, and polarity effects.

[0078] The post-reconstruction processing performed by the reconstruction device 1364 can include image filtering and image smoothing, volume rendering processing, and image difference processing as necessary. The image reconstruction process can be implemented using filtered backprojection, iterative image reconstruction methods, or probabilistic image reconstruction methods. The reconstruction device 1364 can use the memory to store, for example, projection data, reconstructed images, calibration data and parameters, and computer programs.

[0079] The reconfiguration device 1364 may include a CPU (processing circuit) implemented as an individual logic gate, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or another complex programmable logic device (CPLD). The implementation form of the FPGA or CPLD may be encoded in VDHL, Verilog, or any other hardware description language, and the code may be stored directly in the electronic memory inside the FPGA or CPLD, or stored as a separate electronic memory. Further, the memory 1362 may be a non-volatile memory such as a ROM, EPROM, EEPROM, or FLASH (registered trademark) memory. The memory 1362 can be a volatile memory such as a static or dynamic RAM, and a processor such as a microcontroller or a microprocessor may be provided to manage the interaction between the electronic memory and the FPGA or CPLD and the memory.

[0080] Alternatively, the CPU within the reconfiguration device 1364 can execute a computer program that includes a set of computer-readable instructions for implementing the functions described herein, and this program is stored in any one of the non-transitory electronic memory and / or hard disk drive, CD, DVD, FLASH (registered trademark) drive, or any other known storage medium described above. Further, the computer-readable instructions may be provided as a utility application, background daemon, or component of an operating system, or a combination thereof, and are executed in conjunction with a processor such as the Xeon (registered trademark) processor of Intel (registered trademark) in the United States, or the Opteron (registered trademark) processor of AMD (registered trademark) in the United States, and an operating system such as Microsoft 10 (registered trademark), UNIX (registered trademark), Solaris (registered trademark), LINUX (registered trademark), MAC-OS (registered trademark) of Apple Inc. (registered trademark), and other operating systems known to those skilled in the art. Further, the CPU may be implemented as a plurality of processors that cooperate in parallel to execute instructions.

[0081] In one implementation, the reconfigured image may be displayed on the display 1366. The display 1366 can be an LCD display, CRT display, plasma display, OLED, LED, or any other display known in the art.

[0082] The memory 1362 can be a hard disk drive, CD-ROM drive, DVD drive, FLASH (registered trademark) drive, RAM, ROM, or any other electronic storage device known in the art.

[0083] Regarding the above embodiments, the following appendices are disclosed as one aspect and selective features of the invention.

[0084] (1) A method for performing cardiac motion compensation in a computed tomography (CT) imaging system, the method comprising: receiving projection data obtained from an imaging object by the CT imaging system; repeatedly reconstructing the received projection data based on an estimated cardiac motion until a predetermined end criterion is met to generate a motion-compensated image of the imaging object; determining a region of interest (ROI) of blood vessels in the generated motion-compensated image; updating the estimated cardiac motion based on an optimization cost function associated with the determined blood vessel ROI; and outputting the generated motion-compensated image as a final reconstructed image of the imaging object.

[0085] (2) The method according to (1), wherein in the first iteration, the estimated cardiac motion is set to a predetermined value.

[0086] (3) The method according to (1), wherein the predetermined end criterion is that a predetermined number of iterations is completed or the optimization cost function reaches a predetermined threshold.

[0087] (4) The method according to (1), further comprising: reconstructing the received projection data to generate a non-motion-compensated image of the imaging object; performing vessel segmentation based on the generated non-motion-compensated image to create a vessel region mask; and the determining step further comprises specifying a region in the generated motion-compensated image as the determined blood vessel ROI based on the created vessel region mask, and the updating step further comprises calculating the cardiac motion to minimize a loss function of the determined blood vessel ROI and updating the estimated cardiac motion to the calculated cardiac motion.

[0088] (5) The step of performing vascular segmentation further includes extracting a cardiovascular skeleton based on the generated motion-compensated-free image, and forming a created vascular region mask by generating a region for each branch of the extracted cardiovascular skeleton, where the region covers the tubular region and the range of motion of the tubular region, and the tubular region means a blood vessel corresponding to a branch of the extracted cardiovascular skeleton, the method according to (4).

[0089] (6) The tubular region is centered in the middle around the branch of the extracted cardiovascular skeleton, has a radius equal to a predetermined value at the starting point of the branch, and along the axial direction of the branch, the radius of the tubular region gradually decreases from the starting point to the ending point of the branch, the method according to (5).

[0090] (7) The step of extracting further includes performing vascular segmentation based on a vascular filtering-based, dynamic growth method-based, or deep learning-based method on the generated motion-compensated-free image to extract the cardiovascular skeleton, the method according to (5).

[0091] (8) The loss function is based on the entropy of the determined vascular ROI, the edge function of the determined vascular ROI, or a combination thereof, the method according to (4).

[0092] (9) The step of determining further includes performing vascular segmentation based on the generated motion-compensated image to create a vascular region mask, and specifying the region in the generated motion-compensated image as the determined vascular ROI based on the created vascular region mask. The step of updating further includes calculating the motion of the heart so as to minimize the loss function of the determined vascular ROI, and updating the estimated motion of the heart to the calculated motion of the heart, the method according to (1).

[0093] (10) The step of performing vascular segmentation further includes extracting a cardiovascular skeleton based on the generated motion-compensated image, and generating a region for each branch of the extracted cardiovascular skeleton to form a created vascular region mask, where the region covers a tubular region and the range of motion of the tubular region, and the tubular region means a blood vessel corresponding to a branch of the extracted cardiovascular skeleton, the method according to (9).

[0094] (11) The tubular region is centered in the middle around the branch of the extracted cardiovascular skeleton and has a radius equal to a predetermined value at the starting point of the branch. Along the axial direction of the branch, the radius of the tubular region gradually decreases from the starting point to the ending point of the branch, the method according to (10).

[0095] (12) The method according to (1) further includes reconstructing the received projection data to generate a non-motion-compensated image of the imaging object, performing vascular segmentation based on the generated non-motion-compensated image to create a vascular template and a vascular region mask, the step of determining further includes specifying a region in the generated motion-compensated image as the determined vascular ROI based on the created vascular region mask, and the step of updating further includes calculating the motion of the heart to maximize the similarity between the determined vascular ROI and the created vascular template, and updating the estimated motion of the heart with the calculated motion of the heart.

[0096] (13) The step of performing vascular segmentation includes forming a created vascular template by extracting a cardiovascular skeleton based on the generated non-motion-compensated image and generating a tubular region for each branch of the extracted cardiovascular skeleton, where the tubular region means a blood vessel corresponding to a branch of the extracted cardiovascular skeleton; forming a vascular region mask created by generating a region for each branch of the extracted cardiovascular skeleton, where the region covers the tubular region and the range of motion of the tubular region. The method according to (12) further includes these steps.

[0097] (14) The tubular region is centered at the branch of the extracted cardiovascular skeleton, has a radius equal to a predetermined value at the starting point of the branch, and along the axial direction of the branch, the radius of the tubular region gradually decreases from the starting point to the ending point of the branch. The method according to (13).

[0098] (15) The step of determining further includes performing vascular segmentation based on the generated motion-compensated image to create a vascular region mask and a vascular template, and specifying a region in the generated motion-compensated image as a determined vascular ROI based on the created vascular region mask. The step of updating further includes calculating the motion of the heart to maximize the similarity between the determined vascular ROI and the created vascular template, and updating the estimated motion of the heart with the calculated motion of the heart. The method according to (1).

[0099] (16) Reconstructing the received projection data to generate an image of the imaging object without motion compensation, and based on the generated image without motion compensation, performing vascular segmentation to create a vascular template and a vascular region mask, further including, the step of determining further includes specifying, based on the created vascular region mask, a region in the generated image with motion compensation as the determined vascular ROI, and the step of updating further includes calculating the motion of the heart so as to minimize an optimization cost function, a weighting term of the optimization cost function including a loss function of the determined vascular ROI, and the similarity between the determined vascular ROI and the created vascular template, and updating the estimated motion of the heart with the calculated motion of the heart, the method according to (1).

[0100] (17) The step of performing vascular segmentation further includes forming the created vascular template by extracting a cardiovascular skeleton based on the generated image without motion compensation and generating a tubular region for each branch of the extracted cardiovascular skeleton, where the tubular region means a blood vessel corresponding to the branch of the extracted cardiovascular skeleton, forming the created vascular template, and forming the created vascular region mask by generating a region for each branch of the extracted cardiovascular skeleton, where the region covers the tubular region and the range of motion of the tubular region, the method according to (16).

[0101] (18) The tubular region is located centrally around the branch of the extracted cardiovascular skeleton, has a radius equal to a predetermined value at the starting point of the branch, and along the axial direction of the branch, the radius of the tubular region gradually decreases from the starting point to the ending point of the branch, the method according to (17).

[0102] (19) The determining step further includes performing vessel segmentation based on the generated motion-compensated image to create a vessel region mask and a vessel template, and identifying, as a determined vessel ROI, a region within the generated motion-compensated image based on the created vessel region mask; the updating step further includes calculating the motion of the heart to minimize an optimization cost function, a weighted term of the optimization cost function including a loss function of the determined vessel ROI, and the similarity between the determined vessel ROI and the created vessel template, and updating the estimated motion of the heart with the calculated motion of the heart. The method according to (1).

[0103] (20) An apparatus for performing cardiac motion compensation in a computed tomography (CT) imaging system including a processing circuit, the processing circuit receiving projection data obtained from an imaging subject by the CT imaging system, and repeatedly reconstructing the received projection data based on an estimated cardiac motion until a predetermined end criterion is met to generate a motion-compensated image of the imaging subject, determining a region of interest (ROI) of vessels within the generated motion-compensated image, updating the estimated cardiac motion based on an optimization cost function associated with the determined vessel ROI, and outputting the generated motion-compensated image as a final reconstructed image of the imaging subject.

[0104] According to at least one of the embodiments described above, motion compensation can be performed.

[0105] Although several embodiments have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, changes, and combinations of the embodiments can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, as well as in the invention described in the claims and the equivalent scope thereof.

Description of Reference Numerals

[0106] 100 device 110 projection data reception circuit 120 image reconstruction circuit 130 blood vessel ROI determination circuit 140 heart motion update circuit 150 reconstructed image output circuit

Claims

1. receiving projection data obtained from an imaging target by an X-ray CT (Computed Tomography) device; iteratively reconstructing the projection data based on estimated cardiac motion to generate a motion compensated image of the imaged object until a predetermined termination criterion is met; determining a vascular Region Of Interest (ROI) within the motion compensated image; updating the cardiac motion based on an optimization cost function associated with the determined vascular ROI; outputting the generated motion compensated image as a final reconstructed image of the imaging object. Image processing methods.

2. The image processing method of claim 1 , wherein in a first iteration, the cardiac motion is set to a predetermined value.

3. The predetermined termination criterion is: The specified number of iterations is completed, or the optimization cost function reaches a predetermined threshold; The image processing method according to claim 1 .

4. reconstructing the projection data to generate a motion-uncompensated image of the imaged object; performing vessel segmentation based on the uncompensated image to generate a vessel region mask; The determining step further includes identifying a region in the motion compensated image as the determined vascular ROI based on the vascular region mask; The updating step includes: Calculating cardiac motion to minimize a loss function of the vascular ROI; updating the estimated cardiac motion to the calculated cardiac motion; Further comprising: The image processing method according to claim 1 .

5. The step of performing vessel segmentation comprises: extracting a cardiovascular skeleton based on the non-motion compensated images; forming the vascular region mask by generating a region for each branch of the cardiovascular skeleton; 5. The image processing method according to claim 4, wherein the regions cover tubular regions and their ranges of motion, the tubular regions meaning vessels corresponding to the branches of the cardiovascular skeleton.

6. 6. The image processing method of claim 5, wherein the extracting step further comprises performing vascular filtering-based, dynamic growing-based, or deep learning-based vascular segmentation on the unmotion compensated image to extract the cardiovascular skeleton.

7. The image processing method of claim 4 , wherein the loss function is based on an entropy of the vascular ROI, an edge function of the vascular ROI, or a combination thereof.

8. The determining step includes: performing vessel segmentation based on the motion compensated image to generate a vessel region mask; and identifying a region in the motion compensated image as the vascular ROI based on the vascular region mask. The updating step includes: Calculating cardiac motion to minimize a loss function of the vascular ROI; updating the estimated cardiac motion to the calculated cardiac motion; Further comprising: The image processing method according to claim 1 .

9. The step of performing vessel segmentation comprises: extracting a cardiovascular skeleton based on the motion compensated images; forming the vascular region mask by generating a region for each branch of the cardiovascular skeleton; 9. The image processing method according to claim 8, wherein the regions cover tubular regions and ranges of motion of the tubular regions, the tubular regions meaning blood vessels corresponding to the branches of the cardiovascular skeleton.

10. reconstructing the projection data to generate a motion-uncompensated image of the imaged object; performing vessel segmentation based on the un-motion compensated image to generate a vessel template and a vessel region mask; The determining step further includes identifying a region in the motion compensated image as the vascular ROI based on the vascular region mask; The updating step includes: calculating cardiac motion to maximize similarity between the vascular ROI and the vascular template; updating the estimated cardiac motion to the calculated cardiac motion; The image processing method of claim 1 , further comprising:

11. The step of performing vessel segmentation comprises: extracting a cardiovascular skeleton based on the non-motion compensated images; forming the vascular template by generating a tubular region for each branch of the cardiovascular skeleton, the tubular region representing a vessel corresponding to the branch of the cardiovascular skeleton; forming the vascular region mask by generating regions for each branch of the cardiovascular skeleton, the regions forming a created vascular region mask covering the tubular regions and ranges of motion of the tubular regions; The image processing method of claim 10, further comprising:

12. The determining step includes: performing vessel segmentation based on the motion compensated image to generate a vessel region mask and a vessel template; and identifying a region in the motion compensated image as the vascular ROI based on the vascular region mask. The updating step includes: calculating cardiac motion to maximize similarity between the vascular ROI and the vascular template; updating the estimated cardiac motion to the calculated cardiac motion. The image processing method according to claim 1 .

13. reconstructing the projection data to generate a motion-uncompensated image of the imaged object; performing vessel segmentation based on the un-motion compensated image to generate a vessel template and a vessel region mask; The determining step further includes identifying a region in the motion compensated image as the vascular ROI based on the vascular region mask; The updating step includes: calculating cardiac motion to minimize an optimization cost function, a weighted term of the optimization cost function including a loss function of the vascular ROI, and a similarity between the vascular ROI and the vascular template; updating the estimated cardiac motion to the calculated cardiac motion. The image processing method according to claim 1 .

14. The step of performing vessel segmentation comprises extracting a cardiovascular skeleton based on the uncompensated image; forming the vascular template by generating a tubular region for each branch of the cardiovascular skeleton, the tubular region representing a vessel corresponding to the branch of the cardiovascular skeleton; forming the vascular region mask by generating regions for each branch of the cardiovascular skeleton, the regions forming a created vascular region mask covering the tubular regions and ranges of motion of the tubular regions; The image processing method of claim 13 further comprising:

15. 15. The image processing method according to claim 5, 9, 11 or 14, wherein the tubular region is centred around the branch of the extracted cardiovascular skeleton and has a radius equal to a predetermined value at the start point of the branch, and along the axial direction of the branch, the radius of the tubular region gradually decreases from the start point to the end point of the branch.

16. The determining step includes: performing vessel segmentation based on the motion compensated image to generate a vessel region mask and a vessel template; and identifying a region in the motion compensated image as the vascular ROI based on the vascular region mask. The updating step includes: calculating cardiac motion to minimize an optimization cost function, a weighted term of the optimization cost function including a loss function of the vascular ROI, and a similarity between the vascular ROI and the vascular template; and updating the estimated cardiac motion to the calculated cardiac motion. The image processing method according to claim 1 .

17. receiving projection data obtained from an imaging subject; iteratively reconstructing the projection data based on estimated cardiac motion to generate a motion compensated image of the imaged object until a predetermined termination criterion is met; determining a vascular Region Of Interest (ROI) within the motion compensated image; updating the cardiac motion based on an optimization cost function associated with the vascular ROI; and processing circuitry for outputting the motion compensated image as a final reconstructed image of the imaged object.

Images