X-ray image feature detection and registration systems and methods

Abstract

To provide, generally, a vascular system and a peripheral vascular system imaging and data collection, and imaging, image processing and feature detection relating thereto; and in part, more specifically, methods for detecting the position and size of contrast cloud in an x-ray image, e.g., with respect to a sequence of x-ray images during intravascular imaging.SOLUTION: Also described herein are methods of detecting and extracting, from x-ray images, metallic wires such as guidewires used in coronary artery procedures. Further, methods of registering vascular trees for one or more images, e.g., in sequences of x-ray images, are disclosed. In part, the disclosure relates to processing, tracking and registering angiography images and elements in such images. The registration can be performed relative to images from an intravascular imaging modality, e.g, optical coherence tomography (OCT) or intravascular ultrasound (IVUS).SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 62 / 257,213, filed November 18, 2015. The contents of that application are incorporated herein by reference in their entirety.

[0002] Technical Field In part, the present disclosure generally relates to a method for imaging and collecting data on the vascular system and peripheral vascular system. More particularly, this disclosure relates in part to image feature detection and analysis. . [Background technology]

[0003] X-ray images provide an important diagnostic tool in many disciplines. Cardiologists use X-ray based imaging systems in a variety of procedures and tests. As a form of conventional x-ray imaging, fluoroscopy is commonly used to perform angiographic imaging of blood vessels. Visualization of the arteries during catheterization is a valuable diagnostic tool. Accurate visualization aids in planning and executing catheter placement and stent deployment. As such, achieving accurate visualization requires X-ray imaging and the identification of relevant features and objects. This is a key technical requirement for tracking the object. Challenges can arise that make such accuracy difficult to achieve. In addition, X-ray-based imaging methods have been combined with intravascular imaging using optical coherence tomography and ultrasound probes. When the imaging device is combined with the imaging system, the imaging and mutual alignment problems become even more complex. Various factors related to the various imaging methods and devices used to acquire and guide the position also play a role. These factors should be addressed. This can create additional problems.

[0004] For example, the injection of a contrast agent to improve the imaging of blood vessels in angiography involves the injection of a contrast agent into an artery. May obscure placed optical, acoustic and other imaging probes. The use of guidewires and catheters in the These factors can interfere with the imaging and analysis tools in different ways. and others for diagnostic intravascular imaging and image display and part of the diagnostic process. In addition, it becomes difficult to perform the image capture, image analysis and alignment. The challenges associated with matching stenotic markers relying on such imaging data and related diagnostic information are This will also have a negative impact on project planning and other procedures.

[0005] The present disclosure addresses these challenges and others. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] U.S. Patent No. 9,351,698 Summary of the Invention [Means for solving the problem]

[0007] The present disclosure generally relates to data collection, imaging, image processing and the like of the vascular and peripheral vascular systems. In part, the present disclosure relates to the field of feature detection, more particularly to For example, for a sequence of X-ray images during intravascular imaging, the location of the contrast cloud in the X-ray images These techniques involve the use of contrast agents in a variety of blood vessels to detect size and function. Introduced into blood vessels or other body lumens as part of a vascular or other x-ray-related imaging procedure. Injection or other delivery of contrast agents is useful in angiography or other x-ray procedures. to render arteries radiopaque for imaging procedures related to heart procedures, e.g. In imaging of arteries, a contrast agent is introduced into one or more arteries. For example, for a sequence of X-ray images, the parameters related to the contrast cloud in the X-ray images are The present invention relates to a method for detecting parameters of a given contrast cloud. This may include a position and a size.

[0008] Detect metal wires from X-ray images, such as guidewires used in coronary artery procedures A method for extracting the sigma-like features, such as those in sequences of X-ray images, is also described in this paper. A method for registering vascular trees for multiple images is also disclosed. The present disclosure relates to a method for processing, tracking and registering angiographic images and elements in such images. The registration involves the alignment of the intravascular imaging modality, e.g., optical coherence tomography. Optical coherence tomography (OCT) or intravascular ultrasound (IVUS) This can be performed on images from a real-time ultrasound (AR ultrasound).

[0009] In part, the present disclosure provides a method for detecting contrast clouds on an x-ray image frame and detecting the clouds in the x-ray image. The dopamine-containing regions can be excluded from one or more subsequent X-ray image processing steps. In this way, the accuracy of cross-frame registration is improved. / Identify stable vascular centerlines to increase the accuracy of trans-frame registration between X-ray frames In addition, thanks to contrast cloud detection, radiopaque markers in X-ray images can be easily detected. -detection can be improved from a more stable vascular centerline. The centerline is also referred to as the trace, and conversely, the trace is also referred to as the vessel centerline.

[0010] In part, the present disclosure relates to a method for analyzing a vascular tree or one or more vascular trees across multiple frames. In part, this disclosure relates to a method for registering multiple segments. A method for extracting branching positions from a skeleton generated from a series of frames and a method for extracting the same branching positions from a series of frames In one embodiment, the method includes eliminating false branches. If not removed, such false branch detection is used to generate paths through the vascular tree. This interferes with accurate image registration and other subsequent processing steps. The disclosure describes the branches according to their characteristics, such as take-off angle, arc length position, The present invention relates to a method for expressing an object based on absolute angle, intensity, and size (scale). In this embodiment, the separation angle is measured relative to the parent vessel. A variety of image processing methods can be used as described herein.

[0011] In part, the present disclosure relates to a method for identifying the same anatomical branch extracted from multiple angiograms using a feature space. This paper presents a method for associating data based on clustering among In part, this disclosure includes methods for filtering and completing missing representations. 4. The method of claim 1, further comprising: This article relates to a method for selecting suitable branch clusters that can function. Cross-sectional registration facilitates position continuity between x-ray images and improves position tracking. This allows for improved accuracy while still allowing for stent deployment and other vascular procedures performed using angiography. Allows a frame of reference for intra-procedures.

[0012] In one embodiment, the method further comprises: Inflate the contrast cloud region to provide a safety zone to exclude marker tracking In one embodiment, the method further comprises: In an embodiment, the method further comprises co-registering the contrast imaging frames. In order to preserve elements having a particular desired scale, one or more angiograms are The method includes applying a smoothing filter to the normal frame.

[0013] In one embodiment, the method further comprises adaptively thresholding the smoothed image. This removes small scale image elements to produce a binary image. In the binary image, pixels in the contrast cloud region are filtered to a first value or threshold, e.g. In one embodiment, the method further comprises: A contrast cloud neighborhood is selected to include a cloud region having the first value, and the neighborhood is divided into or a plurality of angiography frames. and then counting or scoring the amount of pixels in the neighborhood that contain the first value. In one embodiment, the method further comprises adaptively thresholding the pixels that include the first value. and removing pixels from each neighborhood that contain values ​​other than the first value.

[0014] In an embodiment, the method further comprises: and generating one or more image masks to detect such guidewire position and This involves finding stable anatomical anchor points using one or more masks. In one embodiment, the determined anchor point is selected as the distal end point of the vascular centerline. Determining such edges, in one embodiment, improves vascular centerline generation.

[0015] In an embodiment, the method further comprises plotting a plurality of clusters associated with the branches. We then use these clusters to perform cross-frame registration between angiography frames. It involves performing a series of calculations such as branching, bending and vessel centerline end points in each frame. The method is based on arc-length interpolation between vessel segments defined between two anchor points. In certain embodiments, the method further comprises any suitable shortest path or optical path determination algorithm; For example, the Viterbi algorithm can be used to identify multiple anatomically related flexors. Such an algorithm or other method may include grouping the inflection points of the various positions. The likely sequences and associated most likely pathways in the vessel sections were In one embodiment, the cost criterion can be determined based on the change in bending angle or curvature. or curvature analogues, the location of the bend along the vessel centerline, and the difference in the bend angle deviation between successive frames Based on.

[0016] In some embodiments, the method further comprises: Apply an image processing transform, such as a matrix, to one or more frames to generate at least one The method includes removing or modifying features in one or more of the frames. This may include generating a mask for performing feature extraction during a later image processing stage. Here, the feature is a guidewire in an image.

[0017] In part, the present disclosure relates to angiographic image data and a method for detecting the angiographic image data obtained during pullback through a blood vessel. The present invention relates to a processor-based method for co-registering intravascular image data with a vascular endoscopic image. storing a plurality of frames of optical coherence tomography data in a memory; and storing angiography image data. storing a plurality of frames of angiographic image data in a memory; and processing said plurality of frames of angiographic image data. to generate a set of interframe co-registered angiography data; generating vascular centerlines for said plurality of frames of angiographic image data; detecting a probe marker in said plurality of frames of data; and detecting one or more blood vessels. Track the position of the probe marker along the centerline; use the tracked position to identify the vascular The plurality of frames of contrast image data and the plurality of frames of optical coherence tomography data. This involves aligning the frames with each other.

[0018] In an embodiment, the method further comprises: Generate a score indicating the level of confidence in the co-registration between frames of shadow data. In one embodiment, the plurality of frames of angiographic image data and The step of co-registering the plurality of frames of optical coherence tomography data may include performing a computer program. generating a mutual registration table using a registration device, The matching table contains multiple angiography image frames, OCT timestamps for each frame, and a Includes multiple frame-by-frame angiography and optical coherence tomography image frames with time stamps , which contains a score measurement for each co-registered position.

[0019] In an embodiment, the method further comprises: Representing the mutually aligned table and the cone in a user interface. In one embodiment, the method further comprises displaying the information using a computing device. The branches in one or more OCT or angiographic images are aligned according to the co-registration technique. using a user interface configured to display the table and the branch, This includes identifying.

[0020] In one embodiment, the X-ray image registration may include performing a frame alignment with respect to one or more X-ray images. Image registration can be performed as part of the transverse registration of the vascular tree. The method is performed on or with a representation of the arterial segment such as a segment These representations can be expressed as skeletons, vessel or vascular tree coordinates or position systems. A system for generating a geometric model of one or more blood vessels and a path through one or more blood vessels. These various representations can include tracing curves or centerlines. , one or more from another imaging modality, such as IVUS or other intravascular imaging modalities. can be detected, generated or registered with respect to multiple other images.

[0021] <Embodiments Related to Imaging Cloud> In part, the present disclosure provides a method for detecting a contrast cloud and one or more detected contrast clouds. For diagnosis and analysis of angiography image frames containing shadow cloud regions.

[0022] In one aspect, the present disclosure provides a method and apparatus for implementing a method of executing a program comprising: Firmware, hardware, software-based image processing modules or A combination of these installed on your system allows you to perform a specific operation or action. A system of one or more computing devices that can be configured to execute the program. One or more computer programs are executed by a data processing device. The device performs certain operations because it contains instructions that, when executed, cause the device to perform an action. One common aspect is that it can be configured to perform one or more detecting one or more regions of interest in the x-ray image or images; The method includes a device-based method for determining one or more time periods during which the contrast agent is present in the blood vessel. A set of angiographic image frames acquired during a first time period containing storing a plurality of candidate angiograms in one or more angiogram frames in a storage device; detecting a cloud region; and combining the plurality of candidate contrast cloud regions to obtain a fusion image. and generating a fused contrast cloud region, the fused contrast cloud region having a cloud boundary. Other embodiments of this aspect include corresponding computer systems, devices, and one or more A computer program stored on a number of computer storage devices, each of which It is configured to perform the actions of the method.

[0023] In various embodiments, implementations may include one or more of the following features. The method further comprises: filtering the angiographic image data located within the fused contrast cloud region by one or one or more subsequent data processing methods. One or more subsequent data processing methods include generating vascular centerlines. generating one or more vascular centerlines for one or more angiographic image frames; In this method, the one or more subsequent data processing steps include: A frame width of two or more angiographic image frames of said set of angiographic image frames. The method further includes cross-sectional registration of two or more of the sets of angiographic image frames. performing a cross-frame registration of the angiographic image frames.

[0024] In one embodiment, the method further includes generating a set of cross-frame locations. In this method, the one or more subsequent data processing methods include: anchor point extraction; long interpolation; and cross-frame position generation. Further, one or more candidate regions or fused contrast cloud regions may be dilated to form an expanded image. and providing an expanded exclusion zone, and further comprising: and excluding the angiography image data from one or more subsequent data processing methods. The method further comprises co-registering the OCT image data and the plurality of angiography image frames. The method further comprises: storing elements having a particular desired scale. applying a smoothing filter to one or more angiography frames so as to The method further comprises applying a smoothed angiography frame in which small scale features are removed. The step of adaptively thresholding the binary image to generate a binary image. One or more pixels in the dot region are a first value in the binary image.

[0025] In one embodiment, the method further comprises: A contrast cloud neighborhood is selected to include the angiography field region, and the neighborhood is then contrasted with one or more angiography fields. The method further comprises overlaying a pixel in the neighborhood that includes the first value on the frame. The method further comprises the step of counting or scoring the number of cells comprising a first value. adaptively threshold each pixel to remove pixels from each neighborhood that contain values ​​other than the first value. The method further includes generating a mask and using the mask to estimate the proximity of the vascular centerline. The method includes detecting stable anatomical anchor points that act as site endpoints.

[0026] In an embodiment, the method further comprises: Detecting a number of anatomical features; generating clusters of the detected anatomical features; and classifying the clusters. performing cross-frame registration between angiography frames using the raster. Here, a cluster refers to a single anatomical feature extracted from multiple frames. In one embodiment, the anatomical feature is a branch or inflection point in the vascular tree.

[0027] In one embodiment, the method further comprises: determining a likely path through the associated inflection points; A shortest path finding algorithm is used to identify and group multiple anatomically associated inflection points. In this method, the likely paths are determined based on the change in bend angle, curvature, curvature similarity, etc. from a group including the object, the bend position along the centerline, and the difference in the bend angle deviation between successive frames. The method further comprises: to one or more frames to remove features in at least one frame. or modifying the same to perform feature extraction during one or more subsequent image processing steps. generating a mask, where the feature is a guidewire in the image.

[0028] In an embodiment, the method further comprises: The plurality of frames of coherence tomography data are co-registered using a co-registration table. The co-registration table includes: Multiple OCT timestamps per frame, multiple angiography timestamps per frame, and an optical coherence tomography image frame, each of which is aligned with respect to the other. The method further comprises measuring the status of the OCT images and the angiography images. a user interface for displaying a representation of the mutual alignment table and and displaying the result using the computing device. The branch in the OCT image or the angiography image is aligned with the mutual registration table and the Using a user interface configured to display the branch, include.

[0029] In an embodiment, the method further comprises: and displaying the image using a diagnostic system, wherein the plurality of frame transverse positions are A registered angiographic image is selected from the set.

[0030] In the method, one or more steps of the method include: receiving a frame from an angiography system; and one or more electronic inputs for receiving the set and storing the set. a memory device and an input in electrical communication with the one or more memory devices; and one or more computing devices for performing one or more steps of the method. instructions executable by the one or more computing devices to perform the Implemented with a diagnostic system including image filters and image processing software modules The implementation of the described techniques may be implemented in hardware, a method or process, or a computer. may include computer software on a data-accessible medium, or It may be stored on a computer readable medium, such as a non-transitory computer readable medium.

[0031] Embodiments Related to Guidewire Detection and Extraction In part, the present disclosure provides a method for detecting angiographically-induced changes in one or more frames of angiographic data. In part, this disclosure relates to a system and method for detecting wires. Various guidewire segments suitable for acting on angiographic images The present invention relates to diagnostic and image processing methods.

[0032] In one embodiment, the present disclosure provides software that, when run, causes a system to perform an action. The software, firmware, hardware, or combination of these is installed on the system. It can be configured to perform specific behaviors or actions thanks to its One or more computers in a system. A program, when executed by a data processing device, causes the device to perform an action. Constructed to carry out a particular operation or action by virtue of containing instructions One general aspect is to detect one or more points of interest in one or more x-ray images. The method includes a processor-based method for detecting a region obtained during a first time period. storing the set of angiographic image frames in an electronic memory device; detecting a guidewire in one or more frames of the set; detecting an artery segment in one or more frames of the set of frames; and generating a plurality of cross-frame positions for the set of frames. The present processor-based method for detecting one or more regions of interest also includes: One or more frames including a guidewire. One or more regions of interest are detected. The processor-based method also includes detecting one or more of the detected arterial segments. Other embodiments of this aspect include a corresponding computer - A system, an apparatus, and a computer program recorded on one or more computer storage devices and a computer program, each configured to perform the actions of the method. .

[0033] Implementations may include one or more of the following features. The method further includes: The method also includes detecting one or more of the detected regions in one or more frames of the set. The method may further include performing arc length interpolation on the determined arterial segments. Further, the plurality of frame transverse positions are used to generate a frame of the set of angiographic image frames. The method further includes performing cross-frame registration on one or more of the frames. The method further comprises identifying a plurality of anatomical features in a number of regions. The goal is to create a cluster that contains a set of detected anatomical features across multiple frames. where the clusters are generated based on the detected anatomical features at different times. The method further comprises: The method also includes detecting a plurality of candidate contrast classes in one or more angiography frames. The method may also include detecting a wood region in the angiography frame. Regions of one or more frames that are identified as containing contrast cloud regions are then subjected to centerline generation. The method further includes removing one or more of the candidate contrast cloud regions from the process. Define the proximal endpoint of one or more vascular centerlines using one or more selected endpoints. Further, the method includes: detecting a location of the guidewire in one or more blood vessels using the detected end points of the guidewire; The method further comprises: defining a distal end point of the core wire. generating a plurality of vascular centerlines for the angiography image frames. The end points of the guidewire may be used to define the distal end points of one or more vascular centerlines. The method further includes pulling the intravascular data collection probe during the first time period. In one or more frames of the set, the method includes: The method includes detecting the guidewire through the catheter.

[0034] In an embodiment, the method also includes applying a plurality of filters to the angiographic image frames. The method may also include filtering the filtered angiography frames. The method may also include adaptively thresholding using an intensity filter. The intensity filtered angiogram frames are then processed using the adaptive thresholding algorithm. The method may also include generating an intensity filtered frame. Optionally, the method further includes detecting a portion of the guidewire in the frame. The plurality of filters includes a morphological filter and a ridge enhancement filter. Then, the computing device receives one of the one or more frames of the set. or further instructions for performing arc length interpolation on the plurality of detected arterial segments. In one embodiment, the computing device comprises: to perform cross-frame registration of said set of angiographic image frames using Includes further instructions.

[0035] In one embodiment, the computing device is configured to receive one or more angiography frames. and further instructions for detecting a plurality of candidate contrast cloud regions in the image. In one embodiment, the computing device selects from one or more of the candidate contrast cloud regions. Using the selected endpoint, a further step is performed to define the proximal endpoint of one or more vascular centerlines. In an embodiment, the computing device includes instructions for: and further instructions for defining a distal end point of one or more vascular centerlines using the end points of In one embodiment, a computing device comprises: Further instructions are included for generating a plurality of vascular centerlines for the vascular imaging image frame. In one embodiment, the computing device uses the detected guidewire end points to Further instructions for defining a distal end point of the one or more vascular centerlines. In one embodiment, the computing device may include, in one or more frames of the set, Detecting the guidewire comprises applying a plurality of filters to the angiographic image frames. and further comprising: The system may also include adaptively thresholding the frames. The adaptive thresholded angiography frames are then used to apply the intensity filter. The system may also include generating a filtered frame. detecting a guidewire portion in the filtered frame. In one embodiment, the plurality of filters includes a morphological filter and a ridge enhancement filter. Implementation of the described techniques may be achieved in hardware, a method or process, or a computer. The Software may include computer software on any computer-accessible medium.

[0036] One general aspect includes a method for detecting one or more features in an angiographic image. The system includes: one or more memory devices; and a computing device in communication with the memory device, the instructions including instructions executable by a computing device to: The set of angiographic image frames acquired during the first time period is stored in an electronic memory device. detecting a guidewire in one or more frames of said set; detecting arterial segments in one or more frames of said set; generating a plurality of cross-frame positions for the group of frames. The system also includes one or more frames that include the detected guidewire. The system also includes a system for detecting one or more of the detected arterial segments. Other embodiments of this aspect include a corresponding computer system. The computer program, the device and the computer program stored on one or more computer storage devices The method includes a program, each of which is configured to perform the actions of the method.

[0037] Implementations may include one or more of the following features. In some embodiments, The computing device detects one or more of the received signals in one or more frames of the set. Further instructions include for performing arc length interpolation on the detected arterial segments. In an embodiment, the computing device includes further instructions as described herein. The system also uses the plurality of frame transverse positions to determine a position of the angiographic image frame. In some embodiments, performing cross-frame registration of the set. In the method, a computing device detects a plurality of images in one or more angiography frames. Further instructions are included for detecting candidate contrast cloud regions.

[0038] In one embodiment, the computing device is configured to detect one or more of the candidate contrast cloud regions. defines the proximal endpoints of one or more vascular centerlines using multiple selected endpoints In one embodiment, the computing device further comprises instructions for: Further, a method for defining the distal end point of one or more vascular centerlines using the end points of the guidewire is disclosed. In one embodiment, the computing device includes instructions in the set comprising: and generating a plurality of vascular centerlines for a plurality of said angiographic image frames. In one embodiment, the computing device includes instructions for: Further instructions for defining the distal end point of one or more vascular centerlines using the end points. nothing.

[0039] In one embodiment, the computing device is The system further includes instructions for detecting a guidewire in the system. It may also include applying a number of filters to the angiography image frames. The system also includes adaptive thresholding of the filtered angiography frames. The system may also use an intensity filter to adaptively threshold the blood vessels. operating on the angiography frames to generate intensity filtered frames; The system may also include a guide in the intensity filtered frame. In one embodiment, the plurality of filters may include detecting wire portions. The techniques described include morphological and ridge enhancement filters. hardware, method or process or computer program on a computer-accessible medium This document may contain computer software.

[0040] Anatomical Feature Detection Embodiments In part, this disclosure describes an anatomical feature detection and clustering-based validation method. In one aspect, the present disclosure relates to one or more X-ray images. A processor-based method for detecting a region of interest, the method comprising: storing the set of angiographic image frames obtained in an electronic memory device; generating a plurality of vascular centerlines for the angiographic image frames; A binary image of the angiography image frame is generated for each group of the subjects; a skeleton image is extracted from the binary image. This includes generating an image.

[0041] In one embodiment, the method further comprises applying a rib filter to the skeleton image. ) or a temporal filter. In one embodiment, the method includes applying a skeleton In one embodiment, detecting one or more anatomical features in the computed tomography image includes detecting one or more anatomical features in the computed tomography image. In one embodiment, the anatomical feature has multiple branches. In one embodiment, the anatomical feature has multiple bending points. In an embodiment, the method further comprises detecting an anatomical feature in a plurality of frames. generating a cluster comprising a set of detected anatomical features spanning the Here, the clusters are images of detected anatomical features taken in different frames at different times. In one embodiment, the method further comprises: In one embodiment, the method further comprises applying a vessel crossing filter to the image. The image is generated from the vascular centerline. In an embodiment, the method generates a plurality of clusters. where each cluster represents a single anatomical feature extracted from a group of frames. In one embodiment, the method comprises: generating a distance metric of

[0042] In one embodiment, the distance metric is a Euclidean metric. In one embodiment, the method further comprises: detecting an anatomical feature based on whether it is present in two or more angiographic image frames; In one embodiment, the method includes validating the authentication result. The stars are grouped together to form a set of clusters, each with a single representative from each frame of interest. In one embodiment, the method includes generating a combination of one or more clusters. In one embodiment, the clusters are selected based on the arc length standard deviation, normalized The standard deviation of the arc length, the standard deviation of the angle difference, the proximity to other clusters, and the redundancy per frame From the average number of long anatomical feature records and the average number of missing branch records per frame The selected parameter is selected from the group consisting of:

[0043] The present invention relates to various aspects and embodiments, but the various aspects and embodiments disclosed herein are It is understood that the embodiments may be integrated together in whole or in part where appropriate. As such, each embodiment disclosed herein may be modified, as appropriate, for a given implementation. Each aspect can be incorporated to different degrees and in different ways. The steps can be combined without limitation.

[0044] Other features and advantages of the disclosed embodiments will be apparent from the following description and the accompanying drawings. It will be. [Brief description of the drawings]

[0045] The drawings are not necessarily to scale, emphasis generally being placed upon illustrative principles. The present disclosure is not intended to be limiting in any respect, but to be illustrative in all respects. The scope of the present disclosure is defined only by the claims. [Figure 1] 1 is a schematic diagram of an X-ray based imaging system and an intravascular imaging and data collection system according to certain exemplary embodiments of the present disclosure. [Figure 2A] FIG. 1 is a schematic diagram of an angiography image processing and frame tracking system and its components suitable for co-registering an X-ray image with an intravascular image by generating multiple frame cross-positions according to an exemplary embodiment of the present disclosure. [Figure 2B] 1 is an exemplary cross-registration table including various parameters for performing cross-registration between frames and output from performing imaging using an X-ray system during OCT or other intravascular pullback in accordance with an exemplary embodiment of the present disclosure. [Figure 2C] FIG. 1 is a schematic diagram of three OCT image frames and three corresponding angiography image frames acquired during pullback of an OCT probe with a radiopaque marker through an artery, according to an exemplary embodiment of the present disclosure. [Diagram 3] 1 is a flow chart illustrating a method for detecting contrast clouds in one or more X-ray images generated by injecting a contrast agent into a blood vessel, according to an exemplary embodiment of the present disclosure. [Figure 4] 1 is an X-ray image frame of a blood vessel injected with contrast agent after a noise removal process has been applied to the image frame, according to an exemplary embodiment of the present disclosure. [Diagram 5] FIG. 2 illustrates an exemplary binary image of a contrast cloud after adaptive thresholding has been applied to an X-ray image frame of a blood vessel, according to an exemplary embodiment of the present disclosure. [Figure 6] FIG. 13 illustrates an example image resulting from generating a bright pixel score or count or other pixel-based or intensity-based metric in a predefined neighborhood around a contrast cloud, according to an example embodiment of the present disclosure. [Figure 7]1A-1C are diagrams illustrating example images in which potential locations and sizes of contrast clouds are detected from image frames according to certain exemplary embodiments of the present disclosure. [Figure 8] FIG. 13 illustrates an exemplary fused cloud mask from multiple image frames in accordance with certain exemplary embodiments of the present disclosure. [Figure 9] 1 is a flowchart illustrating a method for detecting a metal wire, according to an example embodiment of the present disclosure. [Figure 10] FIG. 1 illustrates an exemplary image enhanced to show an elongated structure, such as a candidate wire, according to certain exemplary embodiments of the present disclosure. [Figure 11] 1A and 1B are diagrams illustrating an example image of a wire and its two end points with a bottom hat filter and an image processing filter such as a ridge filter, according to an example embodiment of the present disclosure, respectively. [Figure 12] FIG. 2 illustrates an exemplary intensity threshold image of a portion of a vascular system with a wire placed within the vessel, according to certain exemplary embodiments of the present disclosure. [Figure 13] FIG. 2 illustrates an exemplary mask suitable for use during an image processing method for detecting a wire, such as an intravascular guidewire, in an X-ray image, in accordance with an exemplary embodiment of the present disclosure. [Figure 14] 1 is a flowchart illustrating a method for aligning vascular tree segments in multiple frames according to one or more detected bends and branches of the vasculature, according to an exemplary embodiment of the present disclosure. [Figure 15A] 1 is an angiography frame pre-processed prior to skeleton image generation, according to an exemplary embodiment of the present disclosure. [Figure 15B] 1 is an original angiography image after application of image processing and data analysis according to certain exemplary embodiments of the present disclosure. [Figure 15C] 1 is a skeleton of one of the major blood vessels in the above image after application of image processing and data analysis, according to certain exemplary embodiments of the present disclosure. [Figure 16] A is a subset of an angiographic image corresponding to a portion of a vascular system according to an exemplary embodiment of the present disclosure, and B is a representation of that portion of the vascular system showing branches as part of a vascular tree and the associated angle β between the vascular branch and its originating (parent) vessel according to an exemplary embodiment of the present disclosure. [Figure 17A] 1 is a plot showing multiple clusters generated from plotting normalized arc length (vertical axis) of branch segments against angle measurements (horizontal axis) in accordance with an exemplary embodiment of the present disclosure. [Figure 17B] FIG. 17B illustrates an angiography image frame depicting various branches (2, 1, and 3) associated with a particular cluster in the plot of FIG. 17A, according to an exemplary embodiment of the present disclosure. [Figure 18] A and B are frames of angiographic image data based on an exemplary embodiment of the present disclosure, showing paths traced through various vessel segments superimposed on the image, with labels identifying junctions or bends identified using the methods and systems described herein. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0046] This disclosure relates to various aspects of X-ray imaging, such as angiography, and its applications in cardiology. In particular, the present disclosure relates to such frame-spanning methods, systems and apparatus. or between such frames, inter-positioning features relating to the frames of angiographic data. The present disclosure relates to registering frames of angiographic data, for example. Improving such co-registration by reducing errors or detecting structures It also relates to various ways of doing so.

[0047] Examples of such error reduction methods and other angiographic or peripheral vasculature imaging enhancements include: Some are discussed in detail in this paper. These embodiments include contrast cloud detection, X-ray imaging, Extracting or identifying wires in a frame of data and, for example, angled Tracking features and devices to the vascular system with respect to branches and bifurcations or guidewires These embodiments relate to aligning through other alignment processes. Reduce errors that can propagate and lead to additional errors and inaccuracies. Ultimately, Such errors are due to the lack of proper cross-frame registration between angiography frames. ame registration), and other imaging modalities, such as angiography frames The error can prevent co-registration with radiopaque mammograms. -tracking probe movement for a probe that includes one or more markers, such as a marker and may interfere with mutual alignment.

[0048] Interventional cardiologists use fluoroscopy in combination with contrast injection for angiographic imaging, etc. The contrast agent spreads through the vascular tree, making it visible on x-ray. In contrast delivery, a time-varying contrast cloud forms near the catheter tip. The positions or points are highly variable, with blob or cloud-like shapes. This highly variable structure obscures the underlying anatomical information and makes tracking and object tracking difficult. or various image processing and computer vision algorithms such as feature detection The cloud may be composed of various lobes or clusters with irregular shapes. As a result, it is possible to combine the various detected regions and Define the overall cloud region by combining the clouds or defining them with an envelope or boundary line. OR or other combination or union operators can be used to define In an embodiment, the contrast cloud is formed by detecting various regions or areas determined to be contrast-containing regions. defines a data exclusion region, defined by the union of neighborhoods.

[0049] The presence of the contrast cloud itself creates unwanted imaging artifacts and errors. Similarly, the presence of one or more guidewires in the artery during imaging may This can lead to errors and misinterpretation of the guidewire. In addition, the tortuous nature of the vascular system can The overlapping nature itself determines the relative position of a given intravascular probe or or other medical devices may be difficult to track where they are located. Also, the flow of angiographic data is very important for both human observers and automated diagnostic systems. Every displayed segment of a blood vessel in the frame has various branches and pathways. Determine how the vascular system responds to twists and turns as well as overlapping sections and aligns Some of the methods and systems described in this paper address these challenges. In view of the above, in part, the present disclosure provides a method for detecting contrast-enhanced x-ray scans, e.g. For example, the present invention relates to a method for detecting the location and spread of a contrast cloud generated during X-ray angiography. do.

[0050] Additionally, catheters, balloons, catheters, and catheter-based catheters are used during various procedural, diagnostic, and intravascular imaging techniques. A variety of wires can be used to guide tents or other devices. As part of the display of information to the user as part of diagnostic and intravascular imaging techniques, The present disclosure relates to wire and / or wire-based imaging of a frame of data, such as an angiography frame. The present invention relates to a method for determining the location of the catheter tip. In addition, this information may be obtained by X-rays and Used to support and enhance user viewing and interpretation of intravascular images The vascular system is a complex network of vessels that trace various branches and trunks. As a result, the present disclosure provides a method for aligning vascular trees in a sequence of X-ray images. In this way, we can see how overlapping arterial branches in x-rays are The branch of the 3D tree corresponds to the heart rate or other phenomena. The branches of the 3D tree are navigated as they change and move from frame to frame in response to the elephant; The above features address factors that can cause alignment errors. This helps improve the accuracy of the cross-frame registration.

[0051] These and other categories of embodiments described herein include optical coherence tomography, ultrasonography, and A variety of imaging and data acquisition systems, including those that work in concert with sonic or other imaging and data acquisition systems. It can be used in X-ray imaging systems. Intravascular imaging techniques are similar to X-ray fluoroscopy or It is a valuable tool that can be used in place of or in combination with other X-ray imaging systems. By looking inside blood vessels, these imaging techniques can provide insight into the state of blood vessels for a given subject. These intravascular images can be used to obtain high-resolution data on the vascular morphology of the blood vessels. The contrast cloud is combined with the transversely registered angiography images acquired between the frames. - Overcoming some of the challenges of noise, overlapping branches and guidewire artifacts directly improves diagnostic accuracy.

[0052] As a result, acoustic techniques such as optical coherence tomography (OCT) and intravascular ultrasound (IVUS) Other intravascular imaging techniques, such as those described in this paper, include: Diagnose, locate and treat vascular disease during interventions such as IPA procedures or stent placement FIG. 1 illustrates a schematic diagram of a system for implementing one or more embodiments of the present invention. An exemplary system 2 for mounting an X-ray imaging system, such as an angiography system, This shows that it includes 4.

[0053] Data acquisition system 2 is a system for acquiring data on nuclear magnetic resonance, X-ray, computer-aided The present invention also includes a non-invasive imaging system, such as a laminar tomography or other suitable non-invasive imaging technique. As a non-limiting example of such a non-invasive imaging system, for example, cine An angiography system 4 suitable for generating an angiogram is shown. The angiography system 4 may include an X-ray fluoroscopy system. This results in angiographic images, typically in the form of frames of image data. A frame of shadow data is generated. This X-ray imaging is performed using the probe in a pullback procedure. is performed while the blood vessels in the region R of the subject S are examined by angiography as well as, for example, The images are taken using one or more imaging techniques such as OCT or IVUS. Imaging results from the scan (left and right images on display 7) and OCT or IVUS, etc. The intravascular imaging results are shown on the central panel of the display 7. In addition to the above displays, The probe used to collect the intravascular data may be disposable, and the system As part of 2, the patient interface unit (PI) It can be connected to the U.

[0054] The angiography system 4 communicates with an angiography data storage and image management system 12. In one embodiment, system 12 is implemented as a workstation or a server. In one embodiment, data processing relating to the acquired angiographic signals can be performed using , directly on the detector of the angiography system 4. The images from the system 4 are The image data is stored and managed by the imaging data storage and image management system 12. In this form, a subsystem, server or workstation handles the functions of the system 12. In one embodiment, the entire system 4 generates electromagnetic radiation, such as x-rays. also receives such radiation after passing through the subject S. Furthermore, the data processing system The system 12 uses signals from the angiography system 4 to measure a region R of the subject S. In one embodiment, the system 12 and the intravascular system All 18 are part of one integrated system.

[0055] As shown in this particular example, the region of interest R may be a region of vasculature such as a particular blood vessel or a peripheral This region R is a subset of the vasculature imaged using OCT or other intravascular modalities. A catheter-based data collection probe 30 is introduced into the subject 10. The probe 30 is inserted into the lumen of a particular blood vessel, such as a coronary artery. Various types of data collection probes, such as OCT probes, FFR probes, IVUS probes, and other probes suitable for intravascular imaging. The probe 30 typically includes a probe tip, one or more Includes a number of radiopaque markers, fiber optics and torque wires. The tip of the probe is connected to one or more data collection subsystems, such as an optical beam director. , acoustic beam director, pressure detector sensor, other transducer or detector This includes devices and combinations of the above.

[0056] For probes that include an optical beam director, the optical fiber 33 is connected to the beam director. The torque wire is connected to the probe with the optical director. In FIG. 1, the optical fiber 33 is shown without the torque wire surrounding it. Furthermore, the probe 30 is covered by a polymer sheath (not shown) that is part of a catheter. In the context of an OCT system, the sample artefact of the interferometer is The optical fiber 33, which is the optics for the patient interface unit (PIU), is connected to the patient interface unit (PIU) as shown in the figure. Optically coupled.

[0057] The patient interface unit PIU receives the end of the probe 30 and connects it to an optical Typically, the data acquisition device includes a probe connector suitable for coupling to a The lobe 30 is disposable. The PIU is based on the type of data collection probe used: Suitable joints and elements are included.

[0058] For example, a combination OCT and IVUS data acquisition probe requires both an OCT and an IVUS PIU. The PIU typically removes the torque wire, sheath and It also includes a motor suitable for pulling back an optical fiber 33 disposed therein. Thus, the blood vessels of the subject 10 can be imaged longitudinally or in cross-section. Lobe 30 is used to measure a specific parameter, such as FFR or other pressure measurements. It can also be done.

[0059] In addition, the PIU is connected to one or more intravascular data acquisition systems 18. The data acquisition system 18 may be an OCT system, an IVUS system, another imaging system, or the above. For example, the probe 30 may be an OCT probe. The system 18 in the kit includes a sample arm of an interferometer, a reference arm of an interferometer, and a photodetector. It can include a test diode, a control system and a patient interface unit. Similarly, as another example, in the context of IVUS systems, intravascular data acquisition systems The stem 18 includes an ultrasonic signal generating and processing circuit, a noise filter, a rotatable joint, The device may include a power supply, a motor and an interface unit.

[0060] In one embodiment, the data acquisition system 18 and the angiography system 4 are Method to synchronize the timestamps of video frames and the timestamps of OCT image frames In one embodiment, the shared clock or other timing signal is configured to The angiography system 12 includes various imaging devices, designated 15a, 15b, and 15c. Run image processing and feature detection and other software-based processes. In an embodiment, the angiography system 12 includes a plurality of angiography devices, as indicated by 15a, 15b, and 15c. Run various image processing and feature detection and other software-based processes These processes are the contrast cloud detection process, the feature extraction process, and the associated work Ear detection and feature extraction, Inter-frame registration process, Cross-frame registration process The process may include any of the processes, methods, and steps described herein.

[0061] Generally, the software-based processes 15a, 15b, and 15c are Reduce errors in registration and perform other processes described herein, e.g. Detect features in X-ray images and use or exclude them in subsequent processing steps. In this way, such software processes are designed to flag Thus, contrast clouds can be detected and then flagged. Loud areas are negatively affected by localization and noise in those areas. will no longer be used for the process.

[0062] In one embodiment, the contrast cloud is positioned near the proximal end of the blood vessel imaged by the x-ray. As part of the process of determining the centerline, it is advantageous to select the end points. The contrast cloud location can be used to select the centerline endpoints from a set of candidate endpoints. In one embodiment, the guidewire is It is advantageous to have the central line located near the distal end of the blood vessel imaged by the X-ray. As part of the process of determining the location of the guidewire, the guidewire position is used to select the endpoint. and select a centerline endpoint from a set of candidate endpoints or define a given centerline endpoint. Helps.

[0063] The present disclosure may include one or more computer program products, namely data Computer for execution by a processing device or for controlling the operation of a data processing device One or more modules of computer program instructions encoded on a readable medium The computer-readable medium may be implemented as a machine-readable storage device, A machine-readable storage substrate, a memory device, or a combination of one or more of these The term "data processing device" refers to any device for processing data. This includes devices and machines, such as programmable processors, computing The present invention relates to a processing device, such as a computer or a plurality of processors or computers. In addition to the hardware, it creates the execution environment for the computer program in question. The code that runs the program, such as processor firmware, protocol stacks, databases, etc. A data management system, an operating system, or a combination of one or more of these The code may include a code for configuring the combination.

[0064] Computer programs (programs, software, software applications) (also known as application, script or code) is It can be written in any programming language, including languages ​​supported by other programs, and can be used as a standalone application. Loan programs or modules, components, subroutines or components Distribute the Software in any manner, including as other units suitable for use in a computing environment. A computer program does not necessarily have to be stored in a file system. The program does not support files that are created by other programs or data (e.g. The part of a file that holds one or more scripts stored in an application language document , in a single file dedicated to the program in question, or in multiple coordinated files (e.g. For example, a set of programs that contain one or more modules, subprograms, or portions of code. A computer program can be stored in a single computer. on a single data center, or located at one site or distributed across multiple sites. deployed to run on multiple computers interconnected by a communication network It is possible.

[0065] The processes and logic flows described in this disclosure operate on input data to produce output. one or more computer programs to perform a function by generating can be executed by one or more programmable processors that execute Special purpose logic circuits, such as FPGAs (Field Programmable Gate Arrays) or Processes and logic flows may also be performed by ASICs (Application Specific Integrated Circuits). The device may also be implemented.

[0066] Processors suitable for the execution of a computer program include, for example, general purpose and Special purpose both microprocessors and any kind of digital computer One or more processors. Typically, a processor has read-only memory or It receives instructions and data from random access memory or both. Computer Books The qualitative elements are a processor for executing instructions and a memory for storing instructions and data. One or more memory devices. Generally, a computer stores data in one or more mass storage devices, such as magnetic, magneto-optical or optical disks, for storing A storage device that contains a disk or receives data from such a storage device operatively coupled to transfer data to, or both.

[0067] The computer or computing device may display a graphical user interface. One or more software modules for displaying the user interface The computing device may include a machine-readable medium or other memory that includes a module. The device exchanges data such as monitoring data or other data through the network. A network may be one or more of wired, optical, wireless, or other Other data exchange connections may be included.

[0068] The computing devices or computers are server computers, client computers, user computers, control systems, intravascular or angiographic diagnostic systems, Actions to be taken by the microprocessor or computing device Any computing device that can execute a set of instructions (sequential or otherwise) that specifies Additionally, the term "computing device" may refer to any device, individually or jointly. Executing a set (or sets) of instructions to perform any one of the software features or methods Runs multiple or acts as one of the system components described in this document The term "computer" is also intended to include any collection of computing devices that

[0069] In addition to the invasive and non-invasive image data collection systems and devices of FIG. Various other types of data may be obtained for the region R and other parameters of interest for the subjects. This includes location information, vessel diameter, vessel branching location, pixel intensity, The region and vicinity of the intensity variation as well as other data can be included.

[0070] The data acquisition system 2 collects angiography frames, OCT frames, and vascular One or more displays showing a user interface for the angiography data 7. The relative position of an angiography frame with respect to other angiography frames. Display 7 may also show other controls and features of interest. It is possible.

[0071] The non-invasive image data generated using the angiography image analysis and processing system 12 includes: One or more servers, which may be system 12 or system 18 shown in FIG. The data can be transmitted to a vascular endovascular instrument or workstation, stored, and processed. The image processing system 16 is in electrical communication with the PIU and the image processing subsystem 18. Subsystem 18 tracks the marker positions and synchronizes the intravascular image frames with the X-ray image frames. It includes various software modules for performing mutual alignment between the Intravascular image data, for example, intravascular data generated using a data collection probe 30. The data frame is connected to a data acquisition and processing system 45 coupled to the probe via a PIU 35. The video frame grabber, e.g. the system 12, a computer board configured to capture angiographic image data from the It can be used in a variety of embodiments.

[0072] As part of the overall configuration of steps and process flow 50 as shown in FIG. Thus, a sequence of X-ray images 51 is generated as output of the angiography system 4 and image processing The data is then transmitted to the data collection system 12 for processing and storage. X-rays are obtained when a pullback intravascular probe is inserted through the artery. Each artery is part of the vascular system. These may lead to various junctions or branches, or to one or more branches. Branches and bifurcations can be visualized by a pullback imaging procedure using the artery of interest, e.g., OCT or IVUS. The guidewire may diverge at various angles from the arterial section being simultaneously received. The extraction subsystem and method 53 determines the position of the probe 30 in a given image frame. It acts against the X-ray to eliminate the appearance of the guide wire used to determine the The position and distal end point and other points along the guidewire can be detected. were also evaluated and used as part of other image and data processing and analysis described in this paper. In one embodiment, the term "extract" as used herein may be used as a car point. can be considered to refer to "detecting" or "discriminating" and vice versa.

[0073] As discussed in this article, imaging of such arteries is facilitated by the introduction of radiopaque contrast agents. This creates a contrast cloud that is visible as an area on each frame of the X-ray image. The cloud forms in the vicinity of where the contrast agent is introduced. The system operates on the X-ray image frame to characterize and / or identify the contrast cloud and its region. A contrast cloud detection subsystem and / or method 55 for detecting or detecting a contrast cloud is also included. The operating system may include software modules 15a, 15b, or 15c or 15d or a combination thereof. The extraction module 55 operates on the X-ray image to detect features of the image or to extract a given frame. Transform one or more image attributes of a frame, such as by increasing or decreasing them. The present invention can be implemented using the following software components: Various flowcharts, steps and processes described in this document or elsewhere herein are intended to be illustrative and not restrictive. Intravascular and angiographic systems and other computing devices described in The present invention can be implemented using device and diagnostic control systems and processors.

[0074] A detection and image processing step is performed to identify problematic contrast-containing regions in the angiographic frame. In addition to identifying and excluding, the present disclosure also provides a processing step for generating vascular centerlines. and software-based methods 60. Centerline generation includes removing contrast cloud regions. Once the centerline is generated, the anchor points can be extracted. Using step and / or method 65, any suitable anatomical feature in the x-ray image may be detected. For example, bifurcations and bends can be detected. These features are useful in a variety of diagnostic and In some embodiments, these structures can be used for image processing methods. Once feature extraction is performed, representative anatomical features, e.g., anatomical features, are identified for each frame. Clusters and groups of bifurcations and bends reduce co-registration errors and low confidence scores It can be used to:

[0075] As a result, the phases are aligned to each other or aligned on a cross-frame basis. Between which angiography frames are shown on the display versus the intravascular image frames of the hand The cross-registration and cross-frame registration process is useful for diagnostic review, scanning, and Facilitates tent deployment review and stent planning. As a result, contrast cloud detection, Reducing registration errors through guidewire extraction and detection of branches and bends , achieving precise mutual alignment and allowing the end user to select the correct position for the arterial or stented condition In various embodiments, the alignment is important for diagnosing the condition. This includes frame cross-registration and frame cross-registration, and vice versa. The loudsensor process 55 and the guidewire extraction process 53 define the centerline endpoints. For example, a first centerline end point value C1 and a second centerline end point value C2 can be generated. C2 is generated from contrast cloud detection 55 and guidewire extraction 53, respectively. Contrast cloud detection data and guidewire detection data can be The generation of proximal and distal endpoint values ​​for the centerline using the Reduces the additional level of computation. Because it uses data.

[0076] As shown in FIG. 2A, the arc length interpolation stage and related stages are Step 67 can be performed on any anatomical object that can be detected on an X-ray image. A signature or feature identifies a corresponding vessel segment in every angiographic frame. In one embodiment, the anatomical landmark or feature can be a bifurcation, The curvatures are then compared to identify the corresponding vessel segments in all angiography frames. Includes branch and bend information as one or more of the following: By using arc-length based interpolation The frame crossing position can be generated by the radiation of the probe. Facilitates tracking of linear opaque marker positions across angiography frames The markers transition through various points during the pullback as shown in Figure 2C. For example, these positions are based on anatomical features in an angiography frame. In one embodiment, the location of these Tracking involves the use of radiopaque markers to perform co-registration with intravascular imaging data. This is combined with tracking of the

[0077] FIG. 2B shows the system of FIG. 1 after the angiography frame has undergone cross-frame mutual registration. 1 shows a co-registration table generated by one or more of the systems, i.e. The table in Fig. 2A is the OCT table. and angiography frames. Because they operate on a stem or data sampling clock or time period. The (x,y) positions of the probe markers in the frame are displayed in a table. A score, which is an indication of the reliability of the alignment between the frame and the OCT frame, is also shown. The angio index is the frame number for an angiographic image sequence. FIG. 2C shows a schematic diagram of a marker on a probe moving to different (x,y) spatial positions over time. A simple representation of the angiogram and which OCT frames correspond to the associated angiogram frames. The markers are located in three intravascular image frames (OCT-0, OCT-1, OCT-2) and three vessels. Through positions A, B, and C in each of the angiography frames (Angio-0, Angio-1, and Angio-2). Motion can be aligned frame-to-frame and across frames.

[0078] Methods and Analysis Related to Contrast Cloud Feature Extraction / Detection In part, the present disclosure relates to a method for improving tracking accuracy of proximal vessel end points. The points are typically located close to the contrast cloud, and are determined by their proximity to the contrast cloud. Contrast cloud detection is an important tool for detecting proximal vessel ends in OCT-angiography co-registration. In particular, the presence of a contrast cloud may improve the stabilization of the OCT-angiography. Co-registration and cross-frame registration between X-ray frames For example, the oss-frame registration is This creates uncertainty when determining a centerline with endpoints that maintain their anatomical position.

[0079] During the X-ray-guided procedure, doctors use contrast agents and Using combined X-ray scans. During contrast injection, the contrast cloud provides contrast readings. May form near the contrast-leading catheter. Wood is typically irregular in shape and can be seen in different image frames collected during a scan. The contrast cloud changes shape and size and may obscure or conceal underlying structures. Potentially, various image processing and computer vision algorithms can be used to This can lead to poor algorithm performance. The location and extent of the clouds can be detected using image processing and computer vision algorithms. When applying the rhythm, a refined region of interest is established.

[0080] For other purposes, it may be used as an image landmark. Essentially, those areas Or the detected contrast cloud is flagged as a noisy or inconsistent region. and the position for which can be used for further image and data processing. data and other image information is excluded because such noisy or inconsistent Data from the region may introduce errors that may affect other subsequent image data transformations. This propagates through the transformation, which in turn causes additional errors and alignment inaccuracies. That is why.

[0081] A contrast cloud detector can be used to generate a binary image. Alternatively, in another embodiment, The bright areas are inverted and the dark areas can be used for cloud detection. In an exemplary embodiment, a fused mask is generated from binary images from multiple image frames. This can be done using a variety of techniques, including using pixel-wise OR operations. It can be generated by:

[0082] A post-filtering stage can be used to filter out small components (single or multiple) that are outside the region of interest. These cloud areas or the sum of the cloud areas or The ORed combination of cloud regions defines the region that should be excluded. In these cloud regions, or a subset thereof, marker tracking is not performed. It is also excluded from the frame crossing process, centerline generation and other processes described in this document. ORed combinations can be combined to produce a combined or This is obtained by performing an OR operation on the merged cloud regions. In order to increase the likelihood of correctly excluding areas where shadow clouds are likely to exist, multiple Candidate cloud regions can be combined. These regions, if used, This can be a source of error and marker position uncertainty, with detrimental effects on subsequent processing steps. You will have.

[0083] Thus, in one embodiment, for an X-ray image frame generated using system 4, When tracking the opaque markers on the imaging probe using the The identified regions are excluded, and no marker tracking is performed there. The same exclusion is This also applies to other detection and processing described in the present specification. The contrast cloud can also be excluded from cross-frame analysis. Detect and define an area associated with the contrast cloud near or Anatomical locations on the contrast cloud can be tracked with greater accuracy.

[0084] In addition, the boundaries or edges of the detected contrast clouds are not used in the registration process or in this paper. Which frame to use to start other processes, such as other processes listed below. This provides a basis for identifying how greater tracking accuracy can be achieved. Improved vessel centerline determination and thus achieved when determining frame transverse position. Improve accuracy, which eliminates imaging artifacts as well as the relative position as shown in FIG. After a mutual registration process has been performed, such as by generating a registration table, Reduce misalignment between tube segments.

[0085] FIG. 3 is a flow chart 10 of a method for detecting contrast clouds for each input image. In step A1, the image is sent to the system for processing for the detection of contrast clouds. In step A2, the image is denoised to produce a smoother image. This step can be optional, but it can improve the image and reduce noise. It can be important for some x-ray images to have better quality images. In this case, step A2 can be skipped. FIG. 4 shows the image of contrast agent 1 after image denoising has been performed. 1 shows an example image of an X-ray image 150 of a blood vessel 170 having 60 regions. Knowing the location of 0 allows for improved consistency and stability of the vascular centerline, which is related to the relative position. This leads to improved accuracy in the alignment and tracking of the markers of the probe.

[0086] In step A3, optionally, strength and shape points are improved (by reducing the number of gaps). A morphological filter is applied to the image to smooth it and improve the cloud uniformity. In one embodiment, steps A2 and A3 are combined in a common denoising step. In one embodiment, steps A2 and A3 are both optical. In step A4, a first adaptive thresholding is applied to the image to produce a binary image. FIG. 5 shows an example image in which adaptive thresholding has been used to produce a binary image 180. As can be seen in Figure 5, the darker areas contain areas of potential contrast clouds. Pixels or portions of the image in FIG. 4 that were bright white areas in the binary image 180 of FIG. 200, 235, 240, and 250. In FIG. 5, the adaptive thresholding The binary image after is shown.

[0087] In step A5, for each image pixel in the binary image generated in step A3, , bright peaks within the vicinity area that resemble the typical size of the contrast cloud that needs to be detected. The number of vertices is counted. The typical neighborhood around the contrast cloud is a disk, rectangular, or rectangular shape. In one embodiment, the adjacent dimensions are less than about 5 mm. In some embodiments, the proximal dimension ranges from about 1 mm to about 4 cm. In the example, the dimension is a diameter, a chord, or a line segment. The bright white peaks in the bright white regions of the image in Figure 5 are included in the predefined neighborhood. An example image 290 resulting from counting the pixels 300, 325, 330, and 335 is shown in FIG. It shows.

[0088] At step A6, adaptive thresholding is performed on each pixel from the image generated at step A5. The adaptive threshold used is as shown in FIG. A threshold related to the size of the neighborhood used to generate the A component filter is used to remove large components from the generated image. In FIG. 6, an image 290 is shown that includes bright areas in a predefined neighborhood. This is what results after the pixel counting step has been performed.

[0089] In step A7, a mass spectrometer such as an image enlargement kernel is used to increase the cloud mask size. A dilation mask step is optionally used using the dilation image processing operator. Example image 35 showing potential locations and sizes of contrast clouds from frames detected 0. The dilated cloud mask 360 is aligned with the edge or boundary 37 0. This is the cloud mask of the contrast cloud in step A10. Cloud masks, individually and together, are used to generate contrast. You can define the region of uncertainty where the clouds lie. Expand the mask ing) reduces the possibility of tracking probes in the cloud region. In an embodiment, cloud regions are detected and defined as exclusion zones.

[0090] In steps B1 to B4, to fuse cloud masks from multiple images, In step B1, a single fused mask is generated. In step B2, cloud masks from multiple images are used together. Incorporating information from the frames pixel by pixel to obtain a merged contrast cloud mask. After obtaining the merge mask, in step B3, another command Use component-based filters to filter out small components or areas of interest. Certain components may be removed during a period of time following the initial delivery of the contrast solution. Given that the cloud will be expanding and distributed across the globe, the use of multiple X-ray frames is a viable option. It is beneficial.

[0091] In step B4, the cloud masks from each frame are fused as shown in Figure 8. Figure 8 shows the fused contrast cloud mask from an X-ray image sequence. 1 shows an example image 160 of the contrast cloud region. This mask was generated to identify the contrast cloud region. This can be applied to the image to differentiate between the two. This can be done by adding a buffer around These identified contrast cloud regions may then be further analyzed using additional These regions can then be stored in memory as areas to be ignored / avoided when performing various image processing. In one embodiment, the fused cloud mask of FIG. 8 is a peak-based cloud mask. It is derived by taking the x-axis OR.

[0092] Methods and Analyses Related to Guidewire and Thin Element Detection Another aspect is the use of X-ray imaging, such as X-ray angiography, which uses thin metal wires. A method is provided for detecting the location of a metal wire used in a procedure of the type described herein. The ear is used to measure physiological conditions in the area around the guidewire or wire. For example, the wire may include a physiological gauge or The wire with the detector can be a pressure wire or a temperature, magnetic, impedance and current wire. and / or gauges for measuring any other condition including but not limited to voltage. The wire may be in the form of a wire.

[0093] The guidewire extraction method described in this paper is stable and consistent across multiple frames. Once the guidewire is detected, Once extracted, the systems and methods described herein use the guidewire position information to A consistent and stable position can be defined. For example, in one embodiment, The guidewire position defined one of the vessel centerline end points in every angiographic frame. In one embodiment, the detected guidewire is selected after detection. The guidewire position is the distal detected guidewire position. Having the guidewire located in the distal portion reduces alignment and frame crossing errors. It can be used to:

[0094] Automatic extraction of wire position and / or wire tip position for various diagnostic procedures For example, the metal wire may contain a physiological measurement gauge, and the The location is used to automatically associate measurements from the gauge with their corresponding anatomical location. In another example, a metal wire is moved in the X-ray scan area. In addition, when the wire is inserted into a specific anatomical position, automatic trajectory extraction can be applied. When the endoscopic catheter is anchored in place, automatic detection of a consistent anatomical location can be achieved using this method. can be achieved.

[0095] FIG. 9 shows a flow chart 500 of a method for detecting wires in an X-ray image. The process flow stages are: x-ray images such as angiogram images or angiogram sequences The process may be performed on multiple images, such as a single image or a cine image. Image smoothing of the X-ray image or image sequence is performed to smooth out elongated structures in the image. This filter is a modified anisotropic diffusion filter, and at each iteration The filter coefficients are calculated by subtracting the original image intensity from the blob and ridge detectors, the Laplacian-Gaussian In one embodiment, the elongated structures are derived from a combination of a vascular , a guidewire, a rib, or one of the other edge-containing elements in the image data. FIG. 10 shows an example image of an X-ray image of a blood vessel, where image smoothing is performed. In FIG. 10, the image is subjected to an image smoothing step. As part of this process, elongated structures are accentuated. The guide wire W is shown on the right and has end points P1 and P2. Two branches B1 and B2 are shown.

[0096] In step 525, a configuration that can eliminate wide structures in the image is used, as shown in FIG. 11A. A morphological filter is applied to the image. In one embodiment, the morphological filter is a bottom hat In one embodiment, the morphological filter is a filter for small scale images such as thin elements. Any filter that is configured or constrained to enhance or select features of the FIG. 11A shows an example image 600 after a morphological filter has been applied. In one embodiment, the morphological filter is a bottom hat filter. The morphological filter is used in a given morphological filter, e.g., a bottom hat filter. Enhancement of dark elongated elements in an image with a typical scale for the structuring element being used. In another embodiment, the morphological filter is a median filter that produces similar results. It can be replaced by a filter.

[0097] In step 530, a ridge enhancement filter or detector is applied, as shown in FIG. 11B. In one embodiment, such a filter is applied to the The filter or detector is applied to the image to enhance the lines in the image, as shown in Figure 11B. A ridge enhancement filter is applied, such as a Frangi filter or other suitable The ridge enhancement filter is implemented using a Hessian filter, a Franck filter, · A filter or other ridge or edge detector may be included.

[0098] In Fig. 11B, a ridge enhancement filter is applied to enhance the ridge structure in the image. 6 shows an example image 605 after the ridge enhancement filter has been applied. It is used to extract thin and elongated features. The ridge enhancement filter output is thresholded. and thin, elongated dark elements that appear as bright pixels in the thresholded image. Generate a binary image containing

[0099] At step 535, a processing step is performed on the input image by performing adaptive thresholding. Thus, the bright areas in the image are rejected. Metal wires are radiopaque and cannot be seen in X-ray images. are visible on the screen as dark strips. To reject image areas with intensity values ​​that are not of interest , an adaptive dichotomy threshold is applied. Thus, in one embodiment, the intensity corresponding to the dark value Bright areas with an intensity greater than some threshold associated with a value or value range are rejected. FIG. 12 shows the result of a process that performs an intensity threshold image processing step. An example image 610 is shown in which the bright areas of the image are rejected.

[0100] In step 540, the ridge enhancement filter output and the adaptive intensity filter output are The merged metal wire mask components are merged using a pixel-wise AND operator. In one embodiment, the angiography image processing software described herein provides The software module and method connects and filters wire fragments detected in the images. In step 550, the wires can be extracted in fragments and the wires are filtered. Other components in the image may be detected that are not related. The fragments are joined together using a combined measurement of the takeoff angle and the distance between the fragments. It can be done.

[0101] In an optional step, the wire is extracted from the surrounding area that may have been interlocked during wire detection. To remove these elements, post-filtering and / or conversion of the component is required. In step 560, a thinning of the components can be performed. The result of the image processing is used to generate a mask of the detected wires. Figure 13 shows 5 shows an example image of the wire mask 620 generated in step 560. 0 is shown with two levels, one intensity level associated with the wire W, The remainder of the mask is a second intensity level, dark or black, that corresponds to the background of the image. The ear W has end points P1 and P2, respectively, as shown. The mask 620 is processed by the thinning process 560 In one embodiment, one or more of the endpoints P1 and P2 are located in the blood vessel. In some embodiments, one or more Guidewire detection is performed to identify a number of distal guidewire end points. One or more distal guidewire end points may be selected to define the vascular centerline end points. can.

[0102] In some embodiments, prior to tracking the vascular centerline distal end point, Then, a wire masking step is performed. The application of a wire mask is performed to mask one or more blood vessels. Tracking distal end points of vascular centerlines on angiographic frames with improved accuracy - Patents.com Assists in identifying contrast cloud regions and tracking such regions with markers. It is easy to have stable, consistent centerline endpoints that are avoided during other image processing steps. Furthermore, having stable and consistent end points improves the accuracy of cross-frame registration. The degree increases.

[0103] Vascular tree alignment Various medical applications rely on the same image captured in different frames during motion and deformation. For example, intravascular imaging pullback and Angiographic imaging of the arteries during stent placement and subsequent stent deployment is one such application. The heart is a fast moving organ and has complex deformations. As a result, such mapping can be difficult to produce because angiography gives a 2D view of a tortuous 3D system. In the 2D view, the components of the vascular tree, branches, guidewires, and intubations can be visualized. The implant, stent and other intravascular imaging devices are superimposed on one another to image the vasculature of the subject. Creating ambiguities and overlapping regions in angiographic images that do not actually exist in the system Various embodiments of the present invention, as disclosed herein, utilize anatomical landmarks. , for performing such a mapping using, for example, bends, anchor points and branches. A method and system are provided.

[0104] In addition, the present disclosure relates to a method for detecting branch extraction on an angiographic image. A method for grouping branches from a x-ray frame is also disclosed. A method for detecting and grouping vascular bends from a frame is disclosed. The motion estimation in the image is also performed by the detection method described in this paper and the frame-by-frame observation method. This can be done using the marks and tracking of their relative movement. By incorporating anchor point detection based on branching and bending across the frame, Methods for improving the accuracy of cross-machine registration can be achieved. The exclusion process can also help improve such cross-frame alignment. .

[0105] In part, the present disclosure relates to a method for detecting an X-ray image on a plurality of X-ray frames, e.g., frames of angiography data. A method for registering vascular trees, vascular tree segments, or other vascular components imaged by In one embodiment, the method includes, for example, As steps or stages in a given alignment, anchor extraction, bending points or splitting In one embodiment, the method includes, for example, extracting a first frame and As a step or stage in a given registration between the second frame, vessel extraction is performed. It can be used.

[0106] As discussed herein, an angiographic image may include one or more branches. The present disclosure provides a method for extracting or otherwise identifying branches on a frame-by-frame or across frames. In one embodiment, multiple X-ray frames are used for image processing and analysis. A method for grouping branches from a frame is described. or image data for extracting or otherwise identifying vascular curvatures across a frame. The processing and analysis steps are described below. In one embodiment, the vessel tortuosity is calculated from multiple X-ray frames. And a method for grouping the branches is described.

[0107] For example, aligning vascular trees extracted from different contrast-enhanced X-ray frames during X-ray angiography. In one embodiment, for each of the vascular branches of interest, a method for In this case, the vascular centerline is known or is treated as known. The detection process can be carried out using various methods described in US Pat. The disclosures of these publications are incorporated herein by reference in their entireties. The alignment method described in (e.g., 3D) as an “anchor” for matching anatomical positions between different frames. Once a set of matching "anchor points" is obtained, the alignment is done by Interpolating between matching locations based on relative geodesic distances measured along the arc length Various distance metrics can be used as appropriate.

[0108] Additionally, "anchor point" fusion is used to generate estimates of three-dimensional cardiac motion and deformation. can be used. Using a pair of angiographic images (2D projections) from the same cardiac phase, A three-dimensional reconstruction of blood vessels can be obtained. These vessels are then visualized in multiple phases of the cardiac cycle. The 3D vascular reconstruction is useful for understanding the heart motion in 3D. The displacements of these anchor points at each view along the 3D vasculature are used to compute the motion estimates in the 3D vasculature. We introduce “anchor point” matching for frame-to-frame alignment of vascular trees. Further details on how to do this are provided below and elsewhere in this document.

[0109] Anatomical feature detection - branching point extraction / detection and grouping FIG. 14 is a diagram of a vascular tree-like cardiac system between the first and second angiography frames. 6 shows an exemplary process flow 630 suitable for aligning the points. The process flow is to detect anatomical features and perform cross-frame / frame Can be used for interframe alignment. Section alignment may be achieved by other anatomical features or landmarks found along the vessel. This can also be achieved by performing frame-to-frame registration of the vascular tree as shown in FIG. It can be used to perform the following:

[0110] In some embodiments, an anatomical feature, such as a bifurcation point, such as a first and second branch of an artery, A process for detecting vascular bifurcations or bends for a given vascular tree and such The associated process of grouping such points is carried out using various data transformation and image processing stages. Initially, the method comprises: A centerline is determined for processing (step C1), such as by user selection or other criteria. .

[0111] The centerline can be determined as described in this article. These x-ray images are An example of such a preprocessed image 640 is shown in FIG. In one embodiment, as a pre-processing step, a skeleton image is generated around each centerline of the object of interest. The result of brightening the peripheral features and darkening the arterial features is shown in FIG. As shown in Fig. 1, the various arterial branches and their associated tortuous and split confluences and their angles are evident. and is detectable.

[0112] 15B and 15C are diagrams showing an original angiography image and an example embodiment of the present disclosure, respectively. After applying image processing and data analysis based on the state of the blood vessel, a skeleton of one of the blood vessels in the image is obtained. The skeleton image of FIG. 15C is the output of stage C3. With continued reference to FIG. 14, the method is divided into two processing paths or categories: In one embodiment, the two processing paths or categories include: The first can relate to an anatomical feature or landmark, one to a bifurcation and one to a posterior bifurcation. The first one concerns the bending point.

[0113] In one embodiment, the branching related portion of the method involves all or part of every frame. A step of detecting splits in the distribution set (step C4) and grouping the splits by clustering The bending-related part of the method includes a looping step (step C5). A step of detecting "inflection" points in all or a subset of the system (step C6) and detecting The method further comprises a step of grouping the obtained bending points (step C7). These groups of branches and bending points are The looping continues by determining the interframe locations of the vascular tree or other vascular structures or subsets thereof. This is used to perform the registration (step C8). Generally, the first anatomical structure and the second Any group of anatomical structures may be grouped or clustered as described herein. The vascular tree or other vascular structures or subsets thereof can then be visualized. This can be used to perform inter-frame registration (step C8).

[0114] Branch detection - a feature related to feature extraction In one embodiment, feature extraction for the branches is performed as described with respect to step C4 of FIG. Performing the process includes various steps. In one embodiment, performing feature extraction, etc. The branch detection is done by applying a shadow removal filter to the original image. In one embodiment, the filtering includes filtering bottom areas with large structuring elements. This is done using the Muhat morphological operation, which reduces the effect of the shadow of the heart on the diaphragm, Further enhance the vessels. The output from the previous stage is filtered using a Hessian-based filter. It can be processed using

[0115] In one embodiment, the output from the Hessian filtering is adaptively thresholded to obtain In one embodiment, a skeletonization algorithm is applied to the binary image to produce a The skeleton image is then obtained by extracting small parts from the skeleton image. This can be improved by eliminating components. In one embodiment, small components In one embodiment, a small component is less than about 5 pixels. In one embodiment, a small component is less than about 15 pixels. In one embodiment, a small component is less than about 20 pixels. In some embodiments, a small component is less than about 25 pixels. Each component is approximately 30 pixels or less.

[0116] In one embodiment, after the generation of the skeleton image and any subsequent enhancements, the method The method includes detecting a set of bifurcations on each frame, e.g., on the centerline of each frame. In one embodiment, the branches are identified as junctions in the skeleton. To obtain accurate results for branching points, false branching-like features such as blood vessel crossings and ribs are used. A feature elimination is required. For that purpose, a series of one or more fields are The filter is applied.

[0117] In one embodiment, the rib cage or individual ribs may be represented in the image; A rib filter is applied to reduce any contributions that may be mistaken as part of the vascular tree structure. The ribs are mostly static compared to the faster moving blood vessels. Any filter configured to reduce static elements from an image for imaging of the imagery. Any filter can also be used here. In one embodiment, a temporal filter is used. Such a filter takes every single pixel in the entire image sequence and performs the It operates by filtering as if it were a 1D signal. In an embodiment, the collection of these signals is filtered by a high pass filter to obtain a static It is desirable to eliminate the background noise. The average image is calculated from multiple frames. The frame of X-ray image data is then subtracted from the frame of interest. When performing analysis and image processing on the It can be viewed as

[0118] In one embodiment, a cross-vessel filter or detector is applied to the x-ray image frames. In a given frame, a vessel crossing occurs such that two adjacent branches separate in opposite directions. The angle of branch separation relative to the main central line may be a factor in the occurrence and progression of such vessel crossings. and related potential for co-registration errors due to crossing vessels. In addition, the vascular crossings are located on different sides of the main centerline and adjacent The branch that satisfies the condition that it has a separation position and an absolute angle difference close to 180° is excluded. This can also be addressed by:

[0119] The process of associating or grouping anatomical features, such as branching, is called clustering. This clustering or grouping process corresponds to step C5 from FIG. In this context, a cluster is a single image extracted from multiple frames. The same branch may occur multiple times, even if in slightly different positions and orientations. If the branch is detected in several frames, the set of representative branches across the frames is Clusters were formed, indicating that the branches were the same, but were imaged at different times in different frames. In some embodiments, one or more links or vectors may be added between clusters. The distance metric between two clusters can be determined by the following equation: It takes into account the differences between the branching features or whatever features are being evaluated and compared. The features were: angle to the main vessel centerline, normalized arc length of the branch, and average length along the branch. Image intensity, branch scale (or width), absolute angle and branch lumen in angiographic images Thus, the above features identify differences and classify a given feature class. This can be used to generate a distance metric for evaluating and comparing tagging behavior. Cut.

[0120] In one embodiment, a branching descriptor space or model is generated using various features. The features are the average image intensity value of the branch (I value), the absolute angle of separation of the branch (A value), and the normalized The feature space contains one or more of the arc lengths (S values) that are used to associate data points in the feature space. In this way, a similarity index and / or a distance index / metric can be used. An example of a suitable metric is the Euclidean metric D(C i ,C j ) can.

[0121] D(C i ,C j )=sqrt((I i -I j ) 2 +(A i -A j ) 2 +(Si -S j ) 2 ) D(C i ,C j ), where I represents the average image intensity, A represents the absolute angle of the separation branch, and S represents the The indices i and j represent the normalized arc lengths of the branches. and j.

[0122] Clusters of branching datasets representing the same branch in multiple frames are detected and The detection / selection process can be performed using feature extraction. The use of the extraction method takes into account the presence of image noise and missing information in the clusters. In one embodiment, a bifurcation is detected over multiple image frames, This set of detections across frames can be clustered together, Feature extraction helps to verify that the branches of a tree are the same across different frames. In some embodiments, this includes filtering out extraneous data such as image noise. Furthermore, feature extraction may involve extracting features into one or more clusters, such as by interpolation or other processes. This may include a step of completing missing information in the

[0123] Clusters of a certain size (large clusters, small or medium sized clusters) ) is selected for the next processing step in one embodiment. A suitably large size for further processing is identified during OCT pullback. Based on the number of representatives compared to the set of angiography frames captured in Once selected, these clusters are used in the cluster merging stage. The process of merging clusters is carried out by creating clusters, each with a single representative from each frame of interest. Generate a set of .

[0124] FIG. 16A shows a subsection of an angiographic image with a bifurcation formed by two branches B1 and B2. FIG. 16B shows a set of vascular trees according to an exemplary embodiment of the present disclosure. A vascular diagram showing the branching as a component and the associated angle β between the branching vessel and its parent vessel. A representation of that portion of the vascular system. As shown in FIG. 16B, the (parent) branch B1 The angle of the branching is shown as angle β in the schematic representation of the branching. The angle β is the angle between branch B1 and branch These are formed from the confluence of B1 and B2 and are indicated by arrows as openings from B1 to B2. The branches of are plotted based on angle and arc length in one embodiment shown in FIG. or grouped in some other way.

[0125] Cluster integration FIG. 17A shows a plot of normalized arc length (vertical axis) against angular measure (horizontal axis). The figure shows the multiple clusters generated from the three clusters of interest: 1, 2, Cluster 1 is to the left of cluster 2 and contains more data. The data points are shown as circular regions. Cluster 3 is located above Cluster 1. There is a larger overlap along the rectangular region that spans the cluster from top to bottom. In one embodiment, a redundancy elimination step is performed.

[0126] For example, for clusters that contain multiple representatives from the same frame, the best fit is A single point that is closest to the target point is selected. The corresponding location for each cluster on a frame of angiography data is shown. It is desirable to have a single cluster representative from the frame. In the absence of representation from one or more of the following sources, the most recent frames may be used as a substitute for the missing information: The interpolated values ​​based on the clustering algorithm are used to complete the cluster. Any type of interpolation (linear, subspaced, etc.) can be used to define the boundaries of such clusters. Other methods (linear, spline-based, curve fitting, etc.) can also be implemented. The elements of the star may, in one embodiment, be arranged to represent such elements as the same elements throughout the images. It is the basis for handling.

[0127] Cluster Selection Each branch cluster is assigned a quality rating that is used for cluster selection. The factors included in the scale are: arc length standard deviation, normalized arc length standard deviation, angle difference standard deviation, etc. proximity to branch clusters (based on the distance between the centroids of different clusters), The average number of redundant branch records per frame, the average number of missing branch records per frame. A weighted average of the various factors may be used to generate a grade. Yes, the clusters with the best grades are finally selected.

[0128] <Vascular tortuosity feature extraction / detection and features> As shown in FIG. 14, the method includes steps C6 and C7. Further details regarding these steps are provided below. A set of anatomical "anchors" The vascular curvature is extracted as a feature based on the location of the vascular curvature. In each frame, the points are defined as points that create corner-like structures with high curvature. In this study, multi-scale curvatures or curvature analogs are extracted from the vessel centerlines described above.

[0129] In an embodiment, the method comprises: To determine the location of each bend, a tracing or shortest path algorithm, such as the Viterbi algorithm, is used. In various embodiments, the anatomical features in a given image frame are Feature extraction is used to detect the markers and other features or landmarks. In this study, we used the bend angle size, arc length, or normalized arc length to measure the bend position along the vessel centerline. A cost criterion based on the difference in the angular deviation between two bends in the same position and in consecutive frames is used. By optimizing, the shortest path features are extracted from all the frames of interest. In an embodiment, the collection of bends and branches across various frames identifies a path. or otherwise used to perform registration between angiography frames.

[0130] After the cost criteria are calculated from multiple starting points, the associated cost for a given solution is calculated. Based on the ranking of the criteria, solutions are extracted. All bending candidates from multiple frames are After the complement is extracted, small bends or bends that display inconsistent positions along the vessel centerline may be detected. A filtering step may be applied to eliminate solutions derived from

[0131] FIG. 18A and B depict two angiographic frames of multiple vascular trees. In frames 700 and 705, three bends were detected from the two X-ray frames. Centerline detection was performed on the frames. The resulting detected centerlines were The white lines are superimposed on the frames. Bend detection is also performed for each X-ray frame. The flexion positions are indicated by white diamonds. In each of the right and left image frames 705, the pixels are numbered 1, 2, and 3 from bottom to top. Three bends are shown. Each vascular centerline traces a path through the artery along bends 3, 2, and 1. The location of a particular bend in all angiographic frames is traced to the anatomical Used as an anchor or reference point to precisely identify a location across the frame can be done.

[0132] Further supporting details on skeleton and vessel centerline generation Additionally, in one embodiment, anatomical feature detection is performed as part of the pre-processing of the angiographic image. In one embodiment, this involves some kind of measurement related to the path taken by the probe in the blood vessel. A skeleton generation process, etc., can be performed to generate a priori information about the species. The generation of line segments through can be used for feature detection. A skeleton is one or more images generated to aid in tracing the blood vessels of the subject being imaged. It is a static object, such as multiple line segments.

[0133] Data that can be used to inform centerline generation and marker tracking A skeleton or line segment vector for generating candidate paths in the blood vessel for the collection probe. The use of a source approach has some drawbacks as opposed to not using such an approach. For example, skeleton-based approaches can be used to Prevents the creation of certain vascular centerlines that pass through the imaging probe / catheter. This can be done.

[0134] Creating a skeleton is a process that involves creating a skeleton of the vessels and tissue being imaged to facilitate centerline generation. Initial candidates for branching and other vascular geometry as a map or framework A method for determining markers and vascular centers is provided by generating a skeleton. To stabilize the line tracing and verify the tracking quality through frames of angiographic image data, This makes it possible to extract points of interest such as branching points and vessel segments.

[0135] In one embodiment, a skeleton is used to detect anatomical features such as branching and vessel geometry. The process of generating the skeleton is implemented during pre-processing of the angiography images. It can be used to detect anatomical features such as major bifurcations and extrapolation points. In addition, the skeleton can be used to detect and generate smooth vessel centerlines. For example, the skeleton can be created using the Viterbi, Dijkstra or central It can be used with other algorithms such as shortest path algorithms to facilitate line generation. The skeleton can be generated based on the preprocessed Hessian image. A user-selected point on the angiographic image that corresponds to the guidewire position is selected from the noise This can be used to reduce the number of iterations and facilitate skeleton generation. This can be implemented by selecting points based on image features.

[0136] In one embodiment, one or more software modules may process the angiography data. is used to generate and track vascular centerlines for a given frame of In this embodiment, the vascular centerline, also referred to herein as the centerline, is measured using optical or acoustic sensors or angiography. Markers associated with other imaging or data collection sensors introduced during data collection. Interactively evaluate each candidate subset of frames of angiographic data for bands. A model or simulation is generated based on the

[0137] In one embodiment, any suitable shortest path algorithm, such as the Viterbi algorithm, or a software module that implements one or more stages of an optical routing algorithm. Dynamic program software modules such as the markers Centerline generation and tracking are typically performed using other algorithms or The centerline tracing is a guide for defining the centerline endpoints. This can be improved by using feature detection such as Dwyer or landmark detection. Defining centerline endpoints improves frame cross registration and centerline determination. Advantageously, the reliability of the system is increased.

[0138] The use of directional arrows, or their absence, in a given drawing may be misleading, as information may flow. It is not intended to limit or require the orientation in which the device may be used. For a given connector, such as an arrow or a line, which is shown as connecting elements, the information is In one or more directions, or in one direction, as appropriate for a given embodiment. The connection may be any type of connection, such as optical, wired, power, wireless or electrical connection. Any suitable data transmission connection may be included.

[0139]

[0023] A method and system for implementing angiography and intravascular data acquisition methods and systems, including, but not limited to, Software Features and Embodiments The following description is of an apparatus hardware suitable for carrying out the disclosed methods described herein. This description is intended to provide an overview of the software and other operating components. It is not intended to limit the environment or scope of applicability of the present disclosure. The hardware and other operating components are suitable as part of the above-described apparatus. This disclosure relates to personal computers, multiprocessor systems, microphones, and the like. processor-based or programmable electronic devices, network PCs, minicomputers It is implemented with other system configurations including PCs, mainframe computers, etc. It is possible.

[0140] Some parts of the detailed description are based on algorithms and data in computer memory. These algorithmic descriptions are presented using symbolic notations of operations on bits. The statements and representations are intended to be illustrative and not restrictive and should not be construed as limiting the scope of the invention. In one embodiment, the algorithm may be, as used herein and generally, The method described in this paper is considered to be a self-consistent sequence of actions that leads to a certain result. The operations performed as steps or otherwise require physical manipulations of physical quantities. Typically, but not necessarily, such quantities are stored, transferred, combined, transformed, compared, and otherwise processed. This takes the form of electrical or magnetic signals that can be manipulated in any way.

[0141] Unless otherwise noted, and as will be clear from the discussion below, throughout this article, we refer to "processing" or " "Computing" or "Calculating" or "Comparing" or "Arc Length Measurement" or "Detecting" or "tracing" or "masking" or "sampling" or "clustering" or feature extraction" or "adaptive thresholding" or "action" or "generation" or "decision" or or "display" or "discover" or "extract" or "filter" or "avoid" or Discussions using terms such as "exclusion" or "interpolation" or "optimization" are computer system or similar electronic computing device, Data represented as physical (electronic) quantities in the system's registers or memory and to operate such computer system memory or registers or other such Such information may be converted into other data represented as physical quantities in storage, transmission or display devices. It is understood that the term refers to the action and process of performing a function.

[0142] The present disclosure, in some embodiments, also relates to apparatus for performing the operations herein. The apparatus may be specially constructed for the required purposes, or may be integrated into a computer Selectively activated or reconfigured by a stored computer program It may be a general-purpose computer.

[0143] The algorithms and displays presented herein may not be implemented in any particular computer or other manner. The present invention is not inherently related to any particular device. Various general-purpose systems may be implemented based on the teachings herein. can be used in conjunction with a program or to perform the required method steps. It may prove useful to build more specialized equipment to do this. The required structure for a variety of these systems will appear from the description below.

[0144] Embodiments of the present invention may be implemented in many different forms, including, but not limited to, the following: Processors (e.g., microprocessors, microcontrollers, etc.) include, but are not limited to: computer for use with a computer controller, digital signal processor, or general purpose computer Computer-programmed logic, programmable logic devices (e.g., field programmed programmable logic for use with FPGAs or other PLDs; Discrete components, integrated circuits (e.g., application specific integrated circuits (ASICs)) or In an exemplary embodiment of the present disclosure, the OC T-probes, FFR probes, angiography systems and other imaging and subject monitoring As part of the processing of data collected using devices and processor-based systems; The whole is implemented as a set of computer program instructions that are executed by a computer. and converting the data into a computer readable format and storing it as such in a computer readable medium, The system is executed by a microprocessor under the control of the operating system. , user interface commands and pullback completion or mutual alignment Demand-based triggers can generate, for example, OCT data and perform the various and other functions listed above. The features and embodiments can be used to understand a processor suitable for performing image processing. is converted into an instruction.

[0145] Computer program logic implementing all or part of the functionality previously described in this document It may be embodied in a variety of forms, including but not limited to: Source code format, computer executable format and various intermediate formats (assembler, , generated by a compiler, linker or locator). The Code is provided for use with various operating systems or operating environments. Various programming languages ​​(e.g. object code, assembly language or Any high level language, e.g. Fortran, C, C++, JAVA or HTML Source code may include a series of computer program instructions implemented in a program. The source code may define and use various data structures and communication messages. It may be in a computer executable form (e.g., via an interpreter) or Alternatively, the source code may be compiled (e.g., via a translator, assembler, or compiler) It may be converted into a computer-executable form.

[0146] Computer programs may be transferred permanently or temporarily in any form (including source code form) The present invention may be implemented in a variety of ways, including but not limited to, by the above-mentioned methods, and may be fixed on a tangible storage medium in any suitable form (e.g., computer executable form, or intermediate form). The tangible storage medium may be a semiconductor memory device (e.g., RAM, ROM, PROM, EEPROM or Flash – programmable random access memory), magnetic memory devices (such as diskettes or fixed disks), optical memory devices (e.g. CD-ROMs), PC cards (e.g. PCMCIA A computer program is a program that uses a USB flash drive (USB flash drive) or other memory device. , in any form of signal that can be transmitted to a computer using any of a variety of communication technologies Such techniques include, but are by no means limited to, analog technology, digital technology, optical technology, wireless technology (e.g. Bluetooth), Includes networking and internetworking technologies. This product together with any accompanying printed or electronic documentation is provided on a removable storage medium and may be distributed in any form (e.g., shrink-wrapped software) Preloaded into a computer system (for example, in the system ROM or on a fixed disk) The information may be stored on a server or bulletin board, or may be transmitted to a communication system (e.g., The content may be distributed via the Internet or the World Wide Web.

[0147] Hardware logic (programmable logic) that implements all or part of the functionality described earlier in this document. (including programmable logic for use with other logic devices) is a replacement for traditional manual The method may be designed using computer-aided design (CAD), hardware, Description languages ​​(e.g. VHDL or AHDL) or PLD programming languages ​​(e.g. PALASM, A Electronically design, capture and simulate using a variety of tools such as BEL or CUPL or may be documented.

[0148] The programmable logic may be fixed, either permanently or temporarily, in a tangible storage medium. The tangible storage medium may be a semiconductor memory device (e.g., RAM, ROM, PROM, EEPROM or Flash - programmable random access memory (RAM), magnetic memory devices (such as diskettes, or fixed disk), optical memory device (e.g. CD-ROM) or other memory device Programmable logic can be implemented using any of a variety of communication technologies. Such technology may be fixed in a signal that can be transmitted to a computer via a computer. This includes, but is not limited to, analog technology, digital technology, optical technology, wireless technology (e.g. Bluetooth), networking and internetworking technologies. The programmable logic is provided with the associated printed or electronic documentation. It may also be distributed on removable storage media (e.g. shrink-wrapped software). software) in a computer system (for example, in the system ROM or on a fixed disk) It may be preloaded or downloaded from a server or bulletin board to the communication system ( For example, it may be distributed over the Internet or the World Wide Web.

[0149] Various examples of suitable processing modules are discussed in more detail below. In the method, a module is a device suitable for performing a particular data processing or data transmission task. It refers to software, hardware or firmware. In some embodiments, modules The rule accepts, transforms, and routes feature extraction and processing instructions or data of various types. A software routine, program or other memory resident application suitable for executing Data may be, for example, angiography data, OCT data, IVUS data, frame data, etc. Cross-section data, pixel coordinates, clusters, clouds, opaque areas, centerlines, shadows, pixels Cells, clusters, distance metrics, intensity patterns, anatomical features, anatomical landmarks, branching , bending and other information of interest described herein.

[0150] The computers and computer systems described herein acquire, process, and Software applications used in communicating, storing and / or communicating with the and a computer-readable medium, such as a memory, for storing the Such memory may be used in conjunction with the computer or computer system to which it is operatively associated. It is understood that a system can be internal, external, remote or local. do.

[0151] Memory may include any means for storing software or other instructions. Well, for example, without limitation, hard disks, optical disks, floppy disks, DVDs (Digital Versatile Disk), CD (Compact Disk), Memory Stick, Flash memory, ROM (read-only memory), RAM (random access memory), DRAM (dynamic Random Access Memory), PROM (Programmable Read Only Memory), EEPROM (Extended Erasable Programmable ROM) ) and / or other similar computer readable media.

[0152] Generally, computer readable media applied in conjunction with the embodiments of the present disclosure described herein. The memory medium is any medium capable of storing instructions that are executed by a programmable device. Where applicable, the methods described herein may include a computer-based memory medium. embodied or executed as instructions stored on a computer-readable memory medium or These instructions may be implemented in C++, C, Java and / or in accordance with embodiments of the present disclosure. Various other types of software programming languages ​​that can be applied to generate instructions The present invention may be implemented in a variety of programming languages ​​such as the above.

[0153] The aspects, embodiments, features and examples of the present disclosure are considered to be illustrative in all respects and are not to be construed as limiting the present disclosure. The scope of the present disclosure is defined only by the claims. Other embodiments, modifications and uses are possible without departing from the spirit and scope of the claimed disclosure. The use of such a method will be apparent to one skilled in the art.

[0154] The use of headings and sections in this application is not intended to limit the disclosure. Each section may be applicable to any aspect, embodiment or feature of the disclosure.

[0155] Throughout this application, compositions are described as having, including, or having particular components. or where a process is described as having, containing or having specific process steps. In this case, the composition of the teachings of the present application may consist essentially of the recited components. It is contemplated that the process of the present teachings may be comprised of any of the components described herein, and Generally, the process comprises or consists of the process steps described. It is also possible that...

[0156] In this application, an element or component refers to the element or component being described. When stated as being included in a list and / or selected from the list, that element or A component may be any one of the elements or components described. or may be selected from the group consisting of two or more of the elements or components described. It should be understood that the compositions, devices and or method elements and / or features may be modified without departing from the spirit and scope of the teachings of this application. These can be combined in a variety of ways, whether explicitly or implicitly stated in this paper. It should be understood that

[0157] Use of the terms "including," "having," or "having" is expressly intended to include, but is not limited to, any term or term used herein, unless expressly stated to the contrary. Unless otherwise stated, the terms should generally be understood as open-ended, without limitation. .

[0158] The use of the singular in this document includes the plural (and vice versa) unless expressly stated otherwise. Further, unless the context clearly dictates otherwise, the singular forms should be interpreted as plural forms. Additionally, when the term "about" is used before a quantitative value, Unless expressly stated, the teachings of this application also include the specific quantitative values ​​themselves. The term "about" refers to a ±10% variation from the nominal value.

[0159] The order of steps or order for performing certain actions may be determined without departing from the spirit or scope of the present teachings. It should be understood that as long as the two or more stages remain the same, it is not important. Steps or actions may be performed simultaneously.

[0160] If a range or list of values ​​is given, between the upper and lower limits of that range or list of values Each intermediate value is considered individually, as if each value were specifically listed in this paper. The present disclosure also includes, as if each of the above and below ranges were included within the scope of the present disclosure. Intermediate smaller ranges, including any of the above, are contemplated and are encompassed within the present disclosure. A list of ranges disclaims any other values ​​or ranges in between, including the upper and lower limits of a given range. It's not something like that.

[0161] Various aspects of the claimed disclosure may be construed as a subset or substage of the techniques disclosed herein. Furthermore, the terminology and The expressions have been used as terms of description and not of limitation, and the use of such terms and expressions The use is not intended to exclude any equivalents of the features shown and described, or portions thereof. It is recognized that various modifications are possible within the scope of the claimed disclosure. What is desired to be protected by the present invention is recited in the appended claims, including all equivalents. These are disclosures that are defined and distinct in accordance with the

[0162] Several aspects will be described. [Aspect 1] A computer that detects one or more regions of interest in one or more x-ray images. 1. A device-based method comprising: The contrast agent is obtained during a first time period that includes one or more time periods during which the contrast agent is in the blood vessel. storing the set of angiographic image frames in an electronic memory device; Detecting multiple candidate contrast cloud regions in one or more angiography frames stages and; combining the plurality of candidate contrast cloud regions to generate a fused contrast cloud region. and wherein the fused contrast cloud region has a cloud boundary. method. [Aspect 2] The angiographic image data located within the fused contrast cloud region is analyzed using one or more of the 2. The method of embodiment 1, further comprising excluding said sample from subsequent data processing methods. [Aspect 3] The one or more subsequent data processing methods include generating vascular centerlines. Method of posting. Aspect 4 Generate one or more vascular centerlines for one or more angiographic image frames. The method of embodiment 3, comprising: Aspect 5 The one or more subsequent data processing methods may include: 3. The method of claim 2, further comprising cross-frame registration of two or more angiographic image frames of the same subject. method. Aspect 6 two or more angiographic image frames of said set of angiographic image frames. The method of embodiment 5, further comprising performing a transverse registration. Aspect 7 7. The method of embodiment 6, further comprising generating a set of cross-frame locations. Aspect 8 The one or more subsequent data processing methods include: anchor point extraction; arc length interpolation; and 3. The method of embodiment 2, comprising a process selected from the group consisting of: cross-frame position generation. Aspect 9 Dilate one or more candidate regions or the fused contrast cloud region to form an expanded providing an exclusion zone, and a blood vessel located within the expanded exclusion zone. and excluding the angiography image data from one or more subsequent data processing methods. 3. The method of embodiment 2, comprising: Aspect 10 Co-registering the OCT image data and the plurality of angiography image frames. The method of embodiment 2, further comprising: Aspect 11 One or more angiographic techniques are used to preserve elements of a particular desired scale. 2. The method of embodiment 1, further comprising applying a smoothing filter to the frame. Aspect 12 Adaptively threshold the smoothed angiography frames with small-scale elements removed. generating a binary image, 12. The method of embodiment 11, wherein the one or more pixels of is a first value. Aspect 13 A contrast cloud neighborhood including a cloud region having the first value in the binary image. and overlaying the vicinity on one or more angiography frames. The method described in 12. Aspect 14 Counting or scoring the number of pixels in said neighborhood that contain said first value. 14. The method of embodiment 13, further comprising: Aspect 15 Adaptively thresholding pixels that include a first value and subtracting from each neighborhood that includes a value other than the first value. The method of embodiment 1, further comprising removing pixels from the Aspect 16 A mask is generated and used to estimate an anatomical feature that serves as the proximal endpoint of the vascular centerline. The method of embodiment 1, further comprising detecting a stable anchor point. Aspect 17 detecting a plurality of anatomical features in said set of angiographic image frames; We generate clusters of anatomical features that are related to the angiography frame. performing cross-frame alignment between the frames, and the clusters are formed from a plurality of frames. 2. The method of embodiment 1, wherein the method refers to a single anatomical feature extracted from the image. Aspect 18 A plurality of anatomically associated inflection points are identified by a probable approximation that passes through the associated inflection points. using a shortest path finding algorithm to identify paths for grouping; The method of embodiment 1. Aspect 19 The likely path includes the bend angle change, the curvature, the curvature analogue, the bend position along the centerline, and and difference in flexion angle deviation between consecutive frames. 20. The method of embodiment 18, wherein the antibody is identified in response to the criterion. Aspect 20 Applying an image processing transform to one or more frames to obtain a result in at least one of the frames. and removing or modifying features in the image and performing feature extraction during one or more subsequent image processing steps. 2. The method of embodiment 1, further comprising generating a mask for execution. Aspect 21 The plurality of frames of angiography image data and the plurality of frames of optical coherence tomography data. and further comprising: registering the frames to each other using a registration table; The co-registration table includes angiography image frames, multiple frame-by-frame OCT time stamps, and , multiple frame-by-frame angiography time stamps and optical coherence tomography image frames and including a score measurement for each co-registered position. Method of posting. Aspect 22 Stent representation in OCT and angiography images, user interface and displaying the mutual registration table and a computing device. 22. The method of embodiment 21, further comprising: Aspect 23 co-registering the branches in one or more OCT images or angiography images; using a user interface configured to display the table and said branches. 22. The method of embodiment 21, further comprising identifying. Aspect 24 Displaying multiple frame transversely registered angiographic images using a diagnostic system and further comprising: The method of embodiment 1, wherein the antibody is selected from the group consisting of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30 Aspect 25 One or more steps of the method include receiving the set from an angiography system. and one or more electronic memory devices for storing said set; one or more memory devices in electrical communication with the input A computing device and a processor for performing one or more steps of the method. or instructions executable by a plurality of computing devices, image filters and images Implemented using a diagnostic system having a processing software module, as described in embodiment 1. How to. Aspect 26 The method of embodiment 17, wherein the anatomical feature is a branch or inflection point in a vascular tree.

Claims

1. 1. A processor-based method comprising: storing in an electronic memory device of a processor-based diagnostic system a set of angiographic image frames of a subject acquired during a first time period, the angiographic image frames including an x-ray imaged region of a blood vessel of the subject; detecting one or more anatomical features on one or more frames of the set of angiographic image frames using the diagnostic system; clustering one or more frames of the set based on the detected one or more anatomical features, where a cluster corresponds to an anatomical feature detected in two or more frames of the set; performing inter-frame registration of angiographic image frames using the cluster of one or more frames; The method includes:

2. 2. The method of claim 1, wherein the one or more anatomical features comprise at least one of a bifurcation, a bend, an angle relative to a vascular centerline, a normalized arc length of the bifurcation, an average image intensity along a bifurcation branch, a bifurcation scale, a bifurcation width, an absolute angle over the one or more frames of the set, or a bifurcation lumen shape.

3. 10. The method of claim 1, wherein detecting the one or more anatomical features further comprises applying, using the diagnostic system, one or more filters to the one or more frames of the set.

4. The method of claim 3 , wherein the one or more filters include at least one of a Hessian-based filter, a shadow removal filter, a rib filter, a temporal filter, or a vessel crossing filter.

5. generating a plurality of centerlines from the imaged region of a blood vessel using the diagnostic system; generating, using the diagnostic system, binary images for a plurality of the angiographic image frames of the set from which centerlines have been generated; generating, for each angiographic frame of the plurality of angiographic image frames, a skeleton image frame using the diagnostic system, the binary image of each such angiographic frame and a centerline generated from the vascular region imaged in each such frame; The method of claim 1 further comprising:

6. The method of claim 1 , wherein the detected one or more anatomical features are detected on one or more frames of a group of skeletal image frames.

7. The method of claim 1 , further comprising generating one or more distance metrics between two or more clusters.

8. The method of claim 7 , wherein the distance metric of the one or more distance measures is a Euclidean metric.

9. The method of claim 7 , further comprising verifying whether the detected one or more anatomical features are present in two or more angiographic image frames.

10. The method of claim 9 , further comprising aggregating the clusters to generate a set of clusters, each having a single representative from each frame of interest.

11. The method of claim 10 , further comprising selecting one or more clusters.

12. 11. The method of claim 10, wherein the clusters are selected based on parameters selected from the group consisting of: standard deviation of branch arc lengths, standard deviation of normalized branch arc lengths, standard deviation of branch angle differences, proximity to other clusters, average number of redundant branch records per frame, and average number of branch records with missing information per frame.

13. aggregating said clusters to generate a set of clusters, each having a single representative from each frame of interest; selecting one or more clusters using the diagnostic system; Further comprising:

2. The method of claim 1, wherein the clusters are selected based on parameters selected from the group consisting of: standard deviation of branch arc lengths, standard deviation of normalized branch arc lengths, standard deviation of branch angle differences, proximity to other clusters, average number of redundant branch records per frame, and average number of branch records with missing information per frame.

14. identifying contrast cloud regions in said set of angiography frames; determining a subset of the set of angiography frames that excludes a region of the set of angiography frames that includes the identified contrast cloud; The method of claim 1 further comprising:

15. one or more memory devices; a computing device in communication with the memory device, the memory device containing instructions executable by the computing device; 1. A system for detecting one or more features in an angiogram, comprising: The instructions cause the computing device to: storing in an electronic memory device of a processor-based diagnostic system a set of angiographic image frames of a subject acquired during a first time period, the angiographic image frames including an x-rayed region of a blood vessel of the subject; detecting one or more anatomical features on one or more frames of the set of angiographic image frames using the diagnostic system; clustering one or more frames of the set based on the detected one or more anatomical features, where a cluster corresponds to an anatomical feature detected in two or more frames of the set; performing inter-frame registration of angiographic image frames using the cluster of one or more frames; A system that executes the above.

16. The system of claim 15 , wherein the computing device further comprises instructions for applying, with the diagnostic system, one or more filters to the one or more frames of the set.

17. The system of claim 15 , wherein the computing device further comprises instructions for generating one or more distance metrics between two or more clusters.

18. The computing device comprises: generating a plurality of centerlines from the imaged region of a blood vessel using the diagnostic system; and generating, using the diagnostic system, binary images for a plurality of the angiographic image frames of the set from which centerlines have been generated; generating, for each angiographic frame of the plurality of angiographic image frames, a skeleton image frame using the diagnostic system, the binary image of each such angiographic frame and a centerline generated from the vascular region imaged in each such frame; The system of claim 15 further comprising instructions for performing:

19. The computing device comprises: aggregating the clusters to generate a set of clusters, each having a single representative from each frame of interest; selecting one or more clusters using the diagnostic system; and further comprising instructions to execute 16. The system of claim 15, wherein the clusters are selected based on parameters selected from the group consisting of: standard deviation of branch arc lengths, standard deviation of normalized branch arc lengths, standard deviation of branch angle differences, proximity to other clusters, average number of redundant branch records per frame, and average number of branch records with missing information per frame.

20. A computer-readable storage medium containing instructions, comprising: The instructions, when executed by one or more processors, cause the one or more processors to: storing in an electronic memory device of a processor-based diagnostic system a set of angiographic image frames of a subject acquired during a first time period, the angiographic image frames including an x-rayed region of a blood vessel of the subject; detecting one or more anatomical features on one or more frames of the set of angiographic image frames using the diagnostic system; clustering one or more frames of the set based on the detected one or more anatomical features, where a cluster corresponds to an anatomical feature detected in two or more frames of the set; performing inter-frame registration of angiographic image frames using the cluster of one or more frames; A computer-readable storage medium for causing the computer to execute the method.

Images