Rigid motion compensation of coronary angiograms

Abstract

This invention relates to processing multiple cardiac images depicting a portion of a coronary artery within an anatomical region of interest. A corresponding set of luminal radius measurements is determined based on each of the multiple cardiac images, and this set of luminal radius measurements includes multiple luminal radius measurements associated with multiple locations, respectively, of the portion of the coronary artery. One or more locations are determined based on the corresponding set of luminal radius measurements, at each of the multiple locations, at least two sets of the corresponding luminal radius measurements having local or relative extrema, such as local maxima or local minima. Rigid motion artifacts in the multiple cardiac images are corrected based on one or more locations.

Description

Technical Field

[0001] Various examples of this disclosure generally relate to angiography. More specifically, various examples relate to processing multiple cardiac images depicting a portion of a coronary artery within an anatomical region of interest to determine one or more vascular landmarks for correcting rigid motion artifacts in the multiple cardiac images. Background Technology

[0002] 3D reconstruction of coronary arteries based on 2D X-ray angiography is crucial for diagnosing cardiovascular disease, guiding interventions, and improving surgical planning. Since angiography typically provides a 2D projection of a 3D structure, advanced computational techniques are used to reconstruct the 3D geometry.

[0003] To reconstruct 3D coronary arteries, at least two angiographic projections are acquired from different angles using methods such as X-ray imaging scanners. However, X-ray imaging scanners typically generate non-simultaneous angiographic projections, which tend to contain several rigid and / or non-rigid motion artifacts (or misalignments), primarily including patient- or imaging device-related movement, respiratory motion, and cardiac motion. To achieve accurate 3D reconstruction of the coronary arteries, these artifacts or misalignments need to be compensated for or corrected before or in conjunction with reconstruction. Summary of the Invention

[0004] Therefore, advanced technologies are needed to correct rigid and / or non-rigid motion artifacts in cardiac images or angiographic projections. Specifically, there is a need for advanced technologies to automatically and accurately correct rigid motion artifacts in multiple cardiac images acquired from different angles. There is also a need for accurate and reliable 3D coronary artery reconstruction for specific patients.

[0005] This need is met by the features of the independent claims. The features of the dependent claims define the embodiments.

[0006] The following describes a technique for correcting rigid motion artifacts in multiple cardiac images depicting a portion of the coronary artery within an anatomical region of interest. The corrected cardiac images can then be used to reconstruct 3D coronary arteries for a patient.

[0007] A computer-implemented method is provided. The method includes acquiring multiple cardiac images. Each of the multiple cardiac images depicts a portion of a coronary artery within an anatomical region of interest. The method also includes determining a corresponding set of luminal radius measurements based on each of the multiple cardiac images, the set of luminal radius measurements comprising multiple luminal radius measurements associated respectively with multiple locations of the portion of the coronary artery. The method also includes determining one or more locations among the multiple locations based on the corresponding set of luminal radius measurements, at each of the one or more locations, at least two sets in the corresponding set of luminal radius measurements having local extrema. The method still includes correcting for rigid motion artifacts in the multiple cardiac images based on one or more locations.

[0008] Based on various examples, multiple heart images can include heart images acquired at different acquisition angles.

[0009] A computing device including a processor and memory is provided. When program code is loaded from memory and executed, the processor is configured to perform a method. The method includes acquiring multiple cardiac images. Each of the multiple cardiac images depicts a portion of a coronary artery within an anatomical region of interest. The method also includes determining a corresponding set of luminal radius measurements based on each of the multiple cardiac images, the set of luminal radius measurements comprising multiple luminal radius measurements associated respectively with multiple locations of the portion of the coronary artery. The method also includes determining one or more locations among the multiple locations based on the corresponding set of luminal radius measurements, at each of the one or more locations, at least two sets in the corresponding set of luminal radius measurements having local extrema. The method still includes correcting for rigid motion artifacts in the multiple cardiac images based on one or more locations.

[0010] An angiography apparatus including a computing device is provided. The computing device includes a processor and a memory. Upon loading and executing program code from the memory, the processor is configured to perform a method. The method includes acquiring multiple cardiac images. Each of the multiple cardiac images depicts a portion of a coronary artery within an anatomical region of interest. The method also includes determining a corresponding set of luminal radius measurements based on each of the multiple cardiac images, the set of luminal radius measurements comprising multiple luminal radius measurements associated respectively with multiple locations of the portion of the coronary artery. The method also includes determining one or more locations among the multiple locations based on the corresponding set of luminal radius measurements, at each of the one or more locations, at least two sets in the corresponding set of luminal radius measurements having local extrema. The method still includes correcting for rigid motion artifacts in the multiple cardiac images based on one or more locations.

[0011] A computer program, computer program product, or computer-readable storage medium including program code is provided. The program code is loadable from memory and executable by a processor. When the program code is loaded from memory and executed, the processor is configured to perform a method. The method includes acquiring a plurality of cardiac images. Each of the plurality of cardiac images depicts a portion of a coronary artery within an anatomical region of interest. The method also includes determining, based on each of the plurality of cardiac images, a corresponding set of luminal radius measurements, the set of luminal radius measurements comprising a plurality of luminal radius measurements associated respectively with a plurality of locations of the portion of the coronary artery. The method also includes, based on the corresponding set of luminal radius measurements, determining one or more locations among a plurality of locations, at each of the one or more locations, at least two sets of the corresponding set of luminal radius measurements having local extrema. The method still includes, based on the one or more locations, correcting for rigid motion artifacts in the plurality of cardiac images.

[0012] It should be understood that, without departing from the scope of the invention, the features mentioned above and those to be explained below may be used not only in the indicated corresponding combinations, but also in other combinations or in isolation. Attached Figure Description

[0013] Figure 1 The illustration schematically depicts aspects of a C-arm machine.

[0014] Figure 2 An exemplary 3D front view of the heart and coronary arteries is schematically illustrated.

[0015] Figure 3 It is a flowchart based on various examples of methods.

[0016] Figure 4 The illustration schematically shows multiple frames of angiography images acquired at different time points.

[0017] Figure 5 An exemplary frame of angiography images according to various examples is schematically illustrated.

[0018] Figure 6 Various exemplary curves, depicting different sets of lumen radius measurements according to various examples, are schematically illustrated.

[0019] Figure 7 It is a block diagram of a computing device based on various examples. Detailed Implementation

[0020] Some examples of this disclosure generally provide multiple circuits or other electrical devices. All references to circuits and other electrical devices and the functions they provide are not intended to be limited to what is illustrated and described herein. While specific labels may be assigned to the various circuits or other electrical devices disclosed, such labels are not intended to limit the scope of operation of the circuits and other electrical devices. Such circuits and other electrical devices may be combined and / or separated from each other in any way based on a desired particular type of electrical implementation. It should be appreciated that any circuit or other electrical device disclosed herein may include any number of microcontrollers, graphics processing units (GPUs), integrated circuits, memory devices (e.g., FLASH, random access memory (RAM), read-only memory (ROM), electrically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or other suitable variations thereof), and software that works in concert to perform one or more of the operations disclosed herein. Additionally, any one or more of the electrical devices may be configured to execute program code embodied in a non-transitory computer-readable medium programmed to perform any number of functions as disclosed.

[0021] In the following description, embodiments of the invention will be described in detail with reference to the accompanying drawings. It should be understood that the following description of the embodiments is not intended to be limiting. The scope of the invention is not intended to be limited by the embodiments or drawings described below, which are considered illustrative only.

[0022] The accompanying drawings are intended to be schematic representations, and the elements illustrated in the drawings are not necessarily shown to scale. Rather, various elements are shown so that their function and general purpose will become apparent to those skilled in the art. Any connection or coupling between functional blocks, devices, components, or other physical or functional units shown in the drawings or described herein may also be achieved through indirect connections or couplings. Coupling between components may also be established via wireless connections. Functional blocks may be implemented using hardware, firmware, software, or a combination of the foregoing.

[0023] The following describes a technique for correcting rigid motion artifacts in multiple cardiac images. By processing multiple cardiac images, each of the multiple cardiac images can depict the same portion of the coronary artery within the anatomical region of interest. For example, each of the multiple cardiac images can depict the same portion of the left or right coronary artery of the same patient.

[0024] Multiple cardiac images can be acquired during or after angiography of a portion of the coronary artery within the anatomical region of interest. This processing can be performed during or after the angiography of the coronary artery portion; that is, real-time processing or post-processing is possible.

[0025] During or after angiography, multiple cardiac images are acquired, each depicting a portion of the coronary artery within an anatomical region of interest, i.e., the same section of the coronary artery. Based on each of the multiple cardiac images, a corresponding set of luminal radius measurements is determined, comprising multiple luminal radius measurements associated with multiple locations of the coronary artery portion, respectively. Based on the corresponding set of luminal radius measurements, one or more locations are determined, at each of the multiple locations, at least two sets of the corresponding luminal radius measurements having local or relative extrema, such as local maxima or local minima. Based on the one or more locations, rigid motion artifacts in the multiple cardiac images are corrected.

[0026] According to various examples, the count of one or more positions is preferably 4 to 6.

[0027] Based on various examples, multiple cardiac images can be acquired during the same cardiac phase. For example, multiple cardiac images can be acquired during the same cardiac phase in different cardiac cycles or at the same time point in different cardiac cycles. For example, multiple cardiac images can be acquired at the end of the diastolic phase in different cardiac cycles. That is, multiple cardiac images encapsulate information associated with substantially the same cardiac motion, such as the corresponding states of the atria, ventricles, and valves.

[0028] Based on various examples, multiple heart images can include heart images acquired at different acquisition angles.

[0029] By processing multiple cardiac images using the techniques disclosed herein, it is possible to compensate for or correct one or more rigid motion artifacts in the multiple cardiac images. Specifically, based on one or more precisely selected or defined vascular landmarks, one or more rigid motion artifacts can be accurately and reliably compensated for or corrected. Then, 3D coronary arteries can be reconstructed based on the corrected multiple cardiac images.

[0030] Generally, the cardiac cycle is the process by which the human heart begins one heartbeat and begins the next. It consists of two phases: a phase called diastole, during which the heart muscle relaxes and refills with blood, followed by a phase called systole, characterized by robust contraction and pumping of blood. After emptying, the heart relaxes and dilates to receive another influx of blood returning from the lungs and those systems before contracting again to pump blood to the lungs and other systems of the body. The heart must fully dilate before it can pump efficiently again in a normally functioning manner. Assuming a healthy heart and a typical heart rate of 70 to 75 beats per minute, each cardiac cycle, or heartbeat, takes approximately 0.8 seconds to complete.

[0031] The events and details of the cardiac cycle can be illustrated using a Wiggers plot. In a Wiggers plot, the X-axis is used to plot time subdivided into cardiac phases, while the Y-axis typically contains the following on a single grid: blood pressure, aortic pressure, ventricular pressure, atrial pressure, ventricular volume, and electrocardiogram (ECG). Therefore, ECG can be used to determine a specific cardiac cycle.

[0032] According to various examples, multiple heart images can be obtained directly from the angiography device or from a database used to store multiple heart images acquired by the angiography device (e.g., a picture archiving and communication system (PACS)).

[0033] Generally, the typical maximum frame rates found on angiography equipment such as C-arm X-ray machines are 8, 15, or 30. This means that 6, 12, or 24 frames can be acquired by a C-arm X-ray machine during a single cardiac cycle of 0.8 seconds.

[0034] According to various examples, angiography equipment may include a C-arm X-ray machine, an intravascular ultrasound machine, an optical coherence tomography (OCT) machine, or a coronary computed tomography (CT) angiography machine.

[0035] As an example of angiography equipment Figure 1 The diagram schematically illustrates aspects of a C-arm machine 800. The C-arm machine 800 can be used for both angiography and fluoroscopy. Figure 1In the C-arm machine 800, the C-arm is in a neutral position P0. The C-arm machine 800 can be manipulated to a rotatable position deviating from the neutral position P0. The angle between the neutral position P0 and the specific rotatable position is called the angiography angle or fluoroscopy angle. The C-arm machine 800 includes a rotatable C-arm 830 on which an X-ray emitting device 831 and an X-ray detection device 832 can be mounted. The C-arm 830, and therefore the X-ray emitting device 831 and the X-ray detection device 832, are positioned centered around a patient surface 840. The X-ray emitting device 831 emits X-rays that can penetrate the patient positioned on the patient surface 840. The X-ray detection device 832 detects the X-rays emitted from the X-ray emitting device 831. When a radio-opaque contrast agent is injected onto the patient's surface 840, and the contrast agent is injected into the patient's blood vessels, some of the X-rays emitted by the X-ray emitting device 831 are absorbed by the radio-opaque contrast agent. This results in the X-ray detection device 832 detecting an image sequence of blood vessels filled with the radio-opaque contrast agent, i.e., an angiographic photograph. The X-ray emitting device 831 and the X-ray detection device 832 can also be collectively referred to as X-ray imaging devices.

[0036] C-arm 830 can be coupled to C-arm rotation unit 820. C-arm rotation unit 820 can be any motorized device configured to rotate C-arm 830 according to angiography angle or fluoroscopy angle. C-arm rotation unit 820 can be attached to and controlled by C-arm control unit 810. C-arm control unit 810 can be any type of circuit capable of controlling C-arm 830. For example, C-arm control unit 810 may include a computing device.

[0037] The C-arm machine 800 may also include a control panel 850 mounted on the side surface of a patient surface support 841. The control panel 850 can be used to control the C-arm 830 to guide a medical specialist to a pathological vessel. Figure 1 No connection between the control panel 850 and the C-arm 830 is shown to simplify the description of the exemplary C-arm machine 800. In some examples, the connection may be wireless. In other examples, the connection may be wired and may be integrated, for example, into the ceiling of the room where the C-arm machine 800 is located.

[0038] The C-arm machine 800 may also include a display 860. The display 860 can be used to display information to medical professionals, such as a real-time fluorescence fluoroscopic image with an overlaid vascular route map image, which includes paths to one or more pathological blood vessels and / or the location of catheter tips. Furthermore, the display 860 can be used to display vascular segmentation data included in the overlaid vascular route map image, including labels for various vascular segments. In some examples, the display 860 may be a touchscreen, which can be used to toggle the display of vascular segmentation data on and off.

[0039] The C-arm robot 800 can connect to a database ( Figure 1 (not shown in the image), such as a PACS located within a hospital's local network, used to store angiographic images and / or measurements of lumen radius associated with multiple locations, respectively, of a portion of the coronary artery.

[0040] Based on various examples, the names of specific segments or the entire coronary artery (e.g., any of those in the "Abbreviations" column in Table 1) can be determined based on the anatomical region of interest from an angiographic examination. For example, such an anatomical region of interest could include the left or right branch of the coronary tree or a specific vascular segment. The anatomical region of interest can be determined based on user input or on the patient's electronic medical record included in a health information system. Additionally or optionally, the anatomical region of interest can be determined based on multiple predefined seed points.

[0041] For example, the portion of the coronary artery within the anatomical region of interest could be one or more segments of the coronary arteries of the heart. Figure 2 An exemplary 3D front view 200 of the heart and coronary arteries is schematically illustrated. Figure 2 The segments of the coronary arteries in the table are based on and numbered according to the vascular segmentation proposed by the American Heart Association (AHA) and revised by the Society for Cardiovascular Computed Tomography (SCCT). Table 1 provides a short description of the corresponding vascular segments and an abbreviation for each segment.

[0042] Table 1: Figure 2 The segmentation of the coronary arteries.

[0043] In addition to the vascular segments listed in Table 1 Figure 2 It also indicates aortic valve 220.

[0044] Figure 3 The illustration shows aspects of processing multiple cardiac images, each depicting a portion of the coronary artery within an anatomical region of interest, i.e., the same section of the coronary artery, such as... Figure 2 The LAD (i.e., a combination of pLAD 206, mLAD 207, and dLAD 208) or RCA (i.e., a combination of pRCA 201, mRCA 202, and dRCA 203) shown are described. Multiple cardiac images are acquired during or after angiography, and thus the processing can be real-time or post-processing. Based on each of the multiple cardiac images, a corresponding set of luminal radius measurements is determined, and this set of luminal radius measurements includes multiple luminal radius measurements associated with multiple locations, respectively, of portions of the coronary artery. Based on the corresponding set of luminal radius measurements, one or more locations are determined, at each of the multiple locations, at least two sets of the corresponding set of luminal radius measurements have local extrema, such as local maxima or local minima. One or more rigid motion artifacts in the multiple cardiac images are corrected based on the one or more locations. Details of method 3000 are described below.

[0045] Box 3100: Obtain multiple cardiac images, each of which depicts a portion of the coronary artery within the anatomical region of interest.

[0046] Based on various examples, multiple heart images can include heart images acquired at different acquisition angles.

[0047] For example, the angiography equipment (e.g., [the image] can be viewed from different angles) Figure 1 A C-arm machine (800) can directly acquire multiple cardiac images. That is, multiple cardiac images can depict different views of the same part of the coronary artery. Additionally or optionally, multiple cardiac images can be acquired at different acquisition angles and at the same time point in different cardiac cycles (such as multiple consecutive cardiac cycles).

[0048] Alternatively, multiple cardiac images can be obtained from a database (e.g., PACS) used to store multiple cardiac images acquired by angiography equipment.

[0049] In the following text, frames from X-ray angiography will be used as examples of cardiac images to explain the technique. That is, a single cardiac image can correspond to a frame from an angiography.

[0050] Figure 4 The illustration schematically depicts multiple frames f1-f10 of angiographic images acquired at different time points t1-t10, with each frame in the multiple frames f1-f10 depicting the same portion C of the coronary artery within the anatomical region of interest.

[0051] For example, a C-arm X-ray machine with a frame rate of 10 can be used to acquire multiple frames f1-f10 during a single cardiac cycle of 1 second. That is, the multiple frames f1-f10 encapsulate information associated with different cardiac movements from multiple time points of a single cardiac cycle, such as the corresponding states of the atria, ventricles, and valves. Such different cardiac movements may be associated with different blood pressure, aortic pressure, ventricular pressure, atrial pressure, ventricular volume, and electrocardiogram.

[0052] To automatically and accurately correct rigid motion artifacts in multiple cardiac images, multiple appropriate frames can be automatically and accurately selected or determined within the frames of angiographic images, in which the entire portion C of the coronary artery is visible.

[0053] According to various examples, obtaining multiple cardiac images may include obtaining angiographic images acquired during angiography of the anatomical region of interest, and selecting multiple cardiac images from frames of the angiographic images based on at least one predefined criterion.

[0054] Such predefined criteria may include at least one of the following: - The coronary arteries are well visible; - Minimal perspective shortening occurs; - No or minimal vascular overlap was observed; - Image quality, especially around the lumen boundary, is good, for example, above the threshold; - Fewer artifacts exist; and - The entire coronary artery is in view.

[0055] Based on various examples, multiple heart images can be selected from frames acquired in different cardiac cycles but at the same time point within the corresponding cardiac cycle.

[0056] For example, the acquired angiographic images may include 100 frames acquired at 10 different acquisition angles over 10 cardiac cycles (e.g., the first to the tenth cardiac cycles), and 10 frames acquired at the same acquisition angle within each cardiac cycle (e.g., the first to the tenth frames). A specific number (e.g., 2, 3, 4, 5, or 6) of cardiac images can be selected from the frames acquired at the 10 different cardiac cycles. Each of such cardiac images may be selected from the same time point within the corresponding cardiac cycle (e.g., as shown in the image). Figure 4 The corresponding frame acquired at any of t1 to t10 shown in the figure.

[0057] Based on various examples, after obtaining multiple cardiac images, method 3000 may also include determining the anatomical region of interest based on multiple predefined seed points.

[0058] For example, proximal seed points may include the tip of the catheter, and distal seed points may include one or more points selected on the main blood vessel or on a lateral branch with significant stenosis.

[0059] Alternatively, the coronary artery portion can be determined based on angular information used to acquire multiple cardiac images.

[0060] According to various examples, method 3000 may also include segmenting the anatomical region of interest across multiple heart images, i.e., separating the anatomical region of interest from the background in each of the multiple heart images.

[0061] Box 3200: Based on each of multiple cardiac images, determine a corresponding set of lumen radius measurements, which includes multiple lumen radius measurements associated with multiple locations of portions of the coronary artery, respectively.

[0062] Based on various examples, any of the existing techniques or algorithms associated with quantitative coronary angiography can be used to determine or calculate the lumen radius measurement. For example, non-patent literature—Reference [1] Garrone, Paolo, et al. "Quantitative coronary angiography in the current era: principles and applications." Journal of interventional cardiology 22.6(2009): 527-536—provides a comprehensive and updated view on quantitative coronary angiography. It is also conceivable that various existing machine learning algorithms or image processing techniques can be used to determine the lumen radius measurement.

[0063] Figure 5 An exemplary frame 1000 of angiographic images according to various examples is schematically illustrated. Frame 1000 can depict, for example... Figure 4 The coronary artery portion C is shown. A centerline 1100 is generated between the start point 1101 and the end point 1124 of coronary artery portion C. Between the start point 1101 and the end point 1124, there are multiple location points 1102-1123. For each location point 1101-1124, the corresponding lumen radius associated with such location point can be measured or determined, and thus a set of lumen radius measurements including the corresponding lumen radius at each location point 1101-1124 can be determined or calculated.

[0064] For example, at location 1107, two lumen radii, r1 and r2, can be determined or measured. The corresponding lumen radius associated with location 1107 can be determined as the average of r1 and r2. Alternatively, the lumen radius associated with location 1107 can be determined or selected as either r1 or r2.

[0065] Similarly, for Figure 4 For each of the multiple frames f1-f10, a corresponding set of lumen radius measurements can be determined or calculated, including the corresponding lumen radius at each of the location points 1101-1124. That is, ten sets s1-s10 of lumen radius measurements can be determined for each of the ten frames f1-f10, and each of these ten sets s1-s10 includes lumen radius measurements associated with location points 1101-1124 respectively.

[0066] To facilitate the determination of the corresponding set of lumen radius measurements s1-s10, the segmentation algorithm can be applied to ten frames f1-f10 respectively to separate the portion C of the coronary artery from the background of the corresponding frame.

[0067] According to various examples, determining the corresponding set of lumen radius measurements may include segmenting a portion C of the coronary artery from a corresponding heart image of multiple heart images, and determining the corresponding set of lumen radius measurements based on the corresponding segmented portion of the coronary artery.

[0068] Box 3300: Based on the corresponding set of lumen radius measurements, determine one or more locations among a plurality of locations, at each of the one or more locations, at least two sets of the corresponding set of lumen radius measurements have local extrema, such as local maximum or local minimum.

[0069] Based on various examples, the corresponding local extrema of the corresponding set of lumen radius measurements can be determined based on the second derivative of the corresponding curve, and each of the corresponding curves depicts the corresponding relationship between the corresponding set of lumen radius measurements and multiple locations of the coronary artery.

[0070] For example, such as Figure 6As shown, after determining the corresponding set of lumen radius measurements (e.g., 10 sets), interpolation methods such as linear interpolation, polynomial interpolation, spline interpolation, and / or mimicry interpolation can be used to determine or plot the corresponding curve (or radius profile) p1-p10 for each set in the set. The x-axis of each curve can indicate the corresponding distance (or 2D projected distance) between the catheter tip and each location point 1101-1124, and the y-axis can indicate the corresponding lumen radius associated with the corresponding location point. The corresponding local extrema of the corresponding set of lumen radius measurements can then be determined based on the local concavity or convexity of the corresponding curves p1-p10. For example, the second derivative can be calculated for each of the location points 1101-1124 for each of the corresponding curves p1-p10. If the corresponding second derivative is less than 0, the corresponding point is a local maximum, and if the corresponding second derivative is greater than 0, the corresponding point is a local minimum.

[0071] Based on various examples, the second derivative can be calculated via central finite difference.

[0072] In practice, the corresponding sets of luminal radius measurements (or radius distribution maps / curves p1-p10) may be misaligned due to vascular deformation caused by cardiac motion. To more accurately correct rigid motion artifacts in multiple cardiac images, the corresponding sets of luminal radius measurements (or radius distribution maps / curves p1-p10) can be distorted before determining one or more locations, at each of the one or more locations, where at least two sets of the corresponding luminal radius measurements have local extrema.

[0073] For example, dynamic time warping algorithms, such as those disclosed in non-patent literature—reference [2] Salvador, Stan, and Philip Chan. "Toward accurate dynamic time warping in linear time and space." Intelligent Data Analysis 11.5 (2007): 561-580—can be used to warp the corresponding set of lumen radius measurements. Alternatively, the optimal subsequence bijection algorithm disclosed in non-patent literature, reference [3] Latecki, Longin Jan, et al. "Optimal subsequence bijection." Seventh IEEE International Conference on Data Mining (ICDM 2007). IEEE, 2007 can also be used. Techniques disclosed in another non-patent literature—reference [4] Müller, Meinard. "Dynamic time warping." Information retrieval for music and motion (2007): 69-84—can still be used.

[0074] Optionally or additionally, method 3000 may further include determining a reference set of lumen radius measurements and aligning each of the corresponding sets of lumen radius measurements with the reference set of lumen radius measurements. The determination of one or more locations may be based on the corresponding aligned sets of lumen radius measurements.

[0075] For example, the reference set of lumen radius measurements can be randomly or arbitrarily selected from the corresponding set of lumen radius measurements. For example, the first set can be selected as the reference set of lumen radius measurements, and other sets can be aligned with the reference set (i.e., the first set) using a dynamic time warp algorithm (e.g., Matlab's dtw function), i.e., executing dtw(first set, second set), dtw(first set, third set), dtw(first set, fourth set), ..., dtw(first set, nth set).

[0076] According to various examples, determining the reference set of lumen radius measurements may include selecting the set of lumen radius measurements with the longest length as the reference set of lumen radius measurements. Optionally or additionally, the set of lumen radius measurements with the longest length may be determined by sorting the respective sets of lumen radius measurements in descending or ascending order based on the respective length of the respective sets.

[0077] For example, each of the corresponding lumen radius measurements can be indicated as a row or column vector, and the length of the corresponding set corresponds to the number of elements or entries of the corresponding vector.

[0078] Optionally or additionally, the alignment may be based on one or more anatomical landmarks within the anatomical region of interest.

[0079] For example, one or more anatomical landmarks may be associated with parts of the coronary artery, such as bifurcation, stenosis, calcification, and so on.

[0080] Box 3400: Correct rigid motion artifacts in multiple cardiac images based on one or more locations.

[0081] For example, to correct rigid motion artifacts in multiple cardiac images, techniques such as those disclosed in non-patent literature—Reference [5] Banerjee, Abhirup, et al. "Point-cloud method for automated 3Dcoronary tree reconstruction from multiple non-simultaneous angiographicprojections." IEEE transactions on medical imaging 39.4 (2019): 1278-1290—can be used. That is, rigid transformations can be estimated for all angiographic images, which minimizes the orthogonal distance between corresponding landmarks in the object domain. One or more locations determined at box 3300 can be used as landmarks or matching landmarks.

[0082] Alternatively or additionally, the techniques disclosed in another non-patent literature—reference [6] Yang, Jian, et al. "Novel approach for 3-D reconstruction of coronary arteries from two uncalibrated angiographic images." IEEE transactions on image processing 18.7 (2009): 1563-1572—may also be used.

[0083] According to various examples, rigid motion artifacts in multiple cardiac images can be corrected by iteratively optimizing the cost function, as disclosed in documents [5] or [6]. The cost function can be formulated using a set of matching landmark points in multiple cardiac images (e.g., one or more locations in a plurality of locations, at each of which at least two sets of the corresponding set of luminal radius measurements have local extrema, such as local maxima or local minima).

[0084] According to various examples, method 3000 may also include determining at least one of the locations of the catheter tip and one or more bifurcation locations of a portion of the coronary artery based on multiple cardiac images. The correction for rigid motion artifacts may also be based on the location of the catheter tip and / or one or more bifurcation locations.

[0085] Depending on various examples, such as those disclosed in document [6], additional camera parameters may be considered additionally or optionally.

[0086] Alternatively or optionally, rigid motion artifacts in multiple cardiac images can be corrected by finding rigid motion between multiple cardiac images using the iterative method disclosed in document [5]. The method disclosed in document [5] iterates between calculating the 3D positions of matching 2D points / landmarks and solving the absolute orientation problem as described in non-patent literature—reference [7], Horn, Berthold KP. "Closed-form solution of absolute orientation using unit quaternions." Josa a4.4 (1987): 629-642.

[0087] The method may include the following steps: In the first step, the midpoint method is used to calculate the 3D position of the matching 2D point, that is, the 3D position is given by the point that is closest to the line formed by the matching point and the camera position.

[0088] In the second step, 3D points affected by rigid body motion are located. These points are projections from the 3D points in step 1 onto the line formed by the camera position in the second view and the 2D points in the second view. The absolute orientation between the 3D points in step 1 and the 3D points affected by rigid body motion in step 2 is the optimal rigid body motion.

[0089] After the second step, update the camera position and 2D point positions using a rigid transformation. Iterate these two steps until convergence.

[0090] According to various examples, if the matched landmark points contain outliers (false matches), techniques such as those disclosed in non-patent literature—reference [8] Fischler, Martin A., and Robert C. Bolles. "Randomsample consensus: a paradigm for model fitting with applications to imageanalysis and automated cartography." Communications of the ACM 24.6 (1981):381-395 or another non-patent literature—reference [9] Barath, Daniel, et al. "MAGSAC++, a fast, reliable and accurate robust estimator." Proceedings of the IEEE / CVFconference on computer vision and pattern recognition. 2020—can be used to handle these outliers to address the absolute orientation problem described in document [7] while eliminating outliers.

[0091] According to this disclosure, one or more of a plurality of locations can be used as matching 2D points or landmarks, wherein at each of the one or more locations, at least two sets of the corresponding lumen radius measurements have local extrema.

[0092] Figure 7 This is a block diagram of a computing device 9000 according to various examples. The computing device 9000 may include a processor 9020, a memory 9030, and an input / output interface 9010. The processor 9020 is configured to load program code from the memory 9030 and execute that program code. When executing the program code, the processor 9020 executes method 3000 for processing multiple cardiac images depicting the same coronary artery of interest (e.g., a portion C of the coronary artery within an anatomical region of interest).

[0093] The computing device 9000 can be embedded in or connected to an angiography device such as a C-arm machine 800. That is, an angiography device including the computing device 9000 can be configured to perform method 3000.

[0094] Refer again Figure 1The C-arm robot 800 may also include a computing device 9000 configured to perform method 3000. The computing device 9000 may be a C-arm control unit 810 and / or a control panel 850. That is, the computing device 9000 may be embedded in or connected to the C-arm robot 800, and therefore the C-arm robot 800 may also be configured to perform method 3000.

[0095] In summary, a technique has been described to facilitate the correction of rigid motion artifacts in multiple cardiac images depicting portions of the coronary arteries within an anatomical region of interest. This technique automatically identifies one or more locations within the coronary artery portion as landmarks or matching 2D points / landmarks for correcting one or more rigid motion artifacts. One or more locations are identified such that at least two sets of corresponding luminal radius measurements at such locations possess local or relative extrema, such as local maxima or local minima. This technique allows for more accurate 3D reconstruction of the coronary arteries.

[0096] As a simple example, the techniques disclosed herein can be used to automatically estimate rigid motion between two non-simultaneous cardiac images (e.g., angiographic views) and thereby correct rigid motion artifacts in the two cardiac images.

[0097] Although this disclosure has been shown and described with respect to certain preferred embodiments, equivalents and modifications will occur to those skilled in the art upon reading and understanding this specification. This disclosure includes all such equivalents and modifications and is limited only by the scope of the appended claims.

[0098] For illustration, X-ray-based angiography explains this disclosure in detail, and the techniques disclosed herein can also be applied to angiography that can be performed using scanning instead of X-rays.

[0099] Furthermore, the method described in this paper can also be used when a portion of the region of interest is not well visible on a certain frame (e.g., due to perspective shortening, overlapping blood vessels, etc.).

[0100] Additionally, the methods described herein can also be used when acquiring multiple cardiac images from different angiographic views or at different acquisition angles. Depending on the acquisition angle, certain regions of interest may be suggested (e.g., LM / pLAD / proximal LCx for spider views). By combining the suggested regions of interest from various angles, a more accurate representation of the entire region of interest can be obtained.

[0101] Reference List [1] Garrone, Paolo, et al. "Quantitative coronary angiography in thecurrent era: principles and applications." Journal of interventionalcardiology 22.6 (2009): 527-536。

[0102] [2]Salvador, Stan, and Philip Chan. "Toward accurate dynamic timewarping in linear time and space." Intelligent Data Analysis 11.5 (2007):561-580。

[0103] [3] Latecki, Longin Jan, et al. "Optimal subsequence bijection."Seventh IEEE International Conference on Data Mining (ICDM 2007). IEEE, 2007。

[0104] [4] Müller, Meinard. "Dynamic time warping." Information retrievalfor music and motion (2007): 69-84。

[0105] [5] Banerjee, Abhirup, et al. "Point-cloud method for automated 3Dcoronary tree reconstruction from multiple non-simultaneous angiographicprojections." IEEE transactions on medical imaging 39.4 (2019): 1278-1290。

[0106] [6] Yang, Jian, et al. "Novel approach for 3-D reconstruction ofcoronary arteries from two uncalibrated angiographic images." Ieeetransactions on image processing 18.7 (2009): 1563-1572。

[0107] [7] Horn, Berthold KP. "Closed-form solution of absolute orientationusing unit quaternions." Josa a 4.4 (1987): 629-642。

[0108] [8] Fischler, Martin A., and Robert C. Bolles. "Random sampleconsensus: a paradigm for model fitting with applications to image analysisand automated cartography." Communications of the ACM 24.6 (1981): 381-395。

[0109] [9] Barath, Daniel, et al. "MAGSAC++, a fast, reliable and accuraterobust estimator." Proceedings of the IEEE / CVF conference on computer visionand pattern recognition. 2020。

Claims

1. A computer-implemented method (3000), comprising: - Obtain (3100) multiple cardiac images, each of which depicts a portion of the coronary artery within an anatomical region of interest; - Based on each of the plurality of cardiac images, determine (3200) a corresponding set of lumen radius measurements, the set of lumen radius measurements including a plurality of lumen radius measurements associated with a plurality of locations of the portion of the coronary artery, respectively; - Based on the corresponding set of lumen radius measurements, determine (3300) one or more of the plurality of locations, at each of the one or more locations, at least two sets of the corresponding set of lumen radius measurements have local extrema; - Based on the one or more locations, correct (3400) rigid motion artifacts in the plurality of cardiac images.

2. The computer-implemented method (3000) according to claim 1 further includes: - Based on the plurality of cardiac images, determine the location of the catheter tip and at least one of one or more bifurcation locations of the portion of the coronary artery; The correction of the rigid motion artifact is also based on the position of the catheter tip and / or one or more bifurcation positions.

3. The computer-implemented method (3000) according to claim 1 or 2, wherein the plurality of cardiac images are acquired during the same cardiac phase.

4. The computer-implemented method (3000) according to any one of the preceding claims, wherein the count of the one or more locations is preferably 4 to 6.

5. The computer-implemented method (3000) according to any one of the preceding claims, wherein obtaining the plurality of cardiac images comprises: - Obtain angiographic images acquired during angiography of the anatomical region of interest; - Select the plurality of cardiac images from the frames of the angiography images based on at least one predefined criterion.

6. The computer-implemented method (3000) according to any one of the preceding claims, wherein the determination of the corresponding set of lumen radius measurements comprises: - Segment the portion of the coronary artery from the corresponding heart images of the plurality of heart images; as well as, - Based on the corresponding segmented portions of the coronary artery, determine the corresponding set of lumen radius measurements.

7. The computer-implemented method (3000) according to any one of the preceding claims further includes: - Determine the reference set of lumen radius measurements; - Align the corresponding set of lumen radius measurements with the reference set of lumen radius measurements; The determination of one or more of the locations is based on a corresponding aligned set of lumen radius measurements.

8. The computer-implemented method (3000) according to claim 7, wherein the determination of the reference set of lumen radius measurement values ​​includes: - Select the set of lumen radius measurements with the longest length as the reference set of the lumen radius measurements.

9. The computer-implemented method (3000) according to any one of the preceding claims further includes: - Based on multiple predefined seed points, the anatomical region of interest is determined.

10. The computer-implemented method (3000) according to any one of the preceding claims, wherein the plurality of cardiac images includes cardiac images acquired at different acquisition angles.

11. A computing device (9000) comprising a processor (9020) and a memory (9030), wherein when loading and executing program code from the memory (9030), the processor (9020) is configured to perform the method (3000) according to any one of the preceding claims.

12. An angiography apparatus (800) comprising the computing device (9000) according to claim 11.

13. A program code executable by a processor, wherein the processor is configured to perform the method (3000) according to any one of claims 1 to 10 when executing the program code.

Images