Method and system for modeling anatomical surfaces
The method employs electroanatomical mapping systems to generate surface points from ultrasound images, construct meshes using Poisson reconstruction, and apply alpha shape algorithms, addressing the challenge of precise anatomical modeling in cardiac procedures.
Patent Information
- Application Number
- PCT/US2026/027165
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-15
- Filing Date
- 2026-05-07
- Publication Date
- 2026-11-19
AI Technical Summary
Existing methods for generating three-dimensional anatomical models from medical images, such as intracardiac echocardiographic images, are inadequate for precise modeling of anatomical surfaces.
A computer-implemented method using electroanatomical mapping systems to generate a set of surface points from ultrasound images, construct an initial mesh using Poisson surface reconstruction, identify interior points, and apply alpha shape algorithms to create a detailed anatomical surface model.
Enables accurate and detailed modeling of anatomical surfaces, including endocardial surfaces, by integrating ultrasound image processing with Poisson surface reconstruction and alpha shape algorithms, enhancing precision in cardiac diagnostic and therapeutic procedures.
Smart Images

Figure US2026027165_19112026_PF_FP_ABST
Abstract
Description
Attorney Docket No. 16043 WOO 1 / 82410- 1319METHOD AND SYSTEM FOR MODELING ANATOMICAL SURFACESCROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of United States provisional application nos. 64 / 055,272, filed 1 May 2026 (“the ‘272 provisional”), and 63 / 806,117, filed 15 May 2025 (“the ‘117 provisional). The ‘272 provisional and the ‘ 117 provisional are hereby incorporated by reference as though fully set forth herein.FIELD
[0002] The present disclosure relates generally to medical procedures, such as cardiac diagnostic and therapeutic procedures, including electrophysiological mapping and cardiac ablation. In particular, the present disclosure relates to generating anatomical surface models.BACKGROUND
[0003] In connection with various cardiac diagnostic and therapeutic procedures, it is known to create a three-dimensional anatomical model of the heart chamber(s) being studied. Those of ordinary skill in the art will recognize that such models are often created by collecting numerous data points from the chamber(s) of interest.
[0004] It would be desirable, however, to be able to generate such three-dimensional anatomical models from medical images, such as intracardiac echocardiographic (ICE) and other ultrasound images.BRIEF SUMMARY
[0005] The instant disclosure provides a computer-implemented method of modeling an anatomical surface bounding an anatomical volume. The method includes receiving a plurality of ultrasound images of the anatomical volume in an electroanatomical mapping system, the electroanatomical mapping system including a display; via the electroanatomical mapping system: generating a set of surface points (z.e., points that are on or very close to (e.g., within about 1 mm of), the anatomical surface being modeled (e.g., the endocardial surface)) from the plurality of ultrasound images; constructing an initial mesh of the anatomical surface from the set of surface points; identifying a set of interior points (i.e., points that are inside the anatomicalAttorney Docket No. 16043 WOO 1 / 82410- 1319surface being modeled (e.g., points within the heart chamber)) using the initial mesh of the anatomical surface; and constructing a model of the anatomical surface from the set of surface points and the set of interior points; and outputting a graphical representation of the model of the anatomical surface to the display.
[0006] The plurality of ultrasound images may include a plurality of two-dimensional ultrasound images.
[0007] The step of generating the set of surface points may include the electroanatomical mapping system: receiving a user input identifying a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; and discretizing the identified blood pool / tissue boundary of the anatomical volume into the set of surface points. Alternatively or addtionally, the step of generating the set of surface points may include the electroanatomical mapping system: segmenting a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; and discretizing the segmented blood pool / tissue boundary of the anatomical volume into the set of surface points.
[0008] The step of constructing the initial mesh of the anatomical surface from the set of surface points can include the electroanatomical mapping system: estimating normals (unit vectors normal to the surface being reconstructed) for the set of surface points; and generating the initial mesh of the anatomical surface using a Poisson surface reconstruction algorithm.
[0009] The step of constructing the initial mesh of the anatomical surface from the set of surface points can further include cropping the initial mesh of the anatomical surface to a bounding box for the set of surface points to generate a cropped initial mesh.
[0010] Still further, the step of constructing the initial mesh of the anatomical surface from the set of surface points can include applying a hole-filling algorithm to the initial mesh of the anatomical surface to generate a filled initial mesh.
[0011] It is contemplated that the method may also include resampling the set of surface points across the initial mesh of the anatomical surface. For instance, the set of surface points may be resampled uniformly across the initial mesh of the anatomical surface.
[0012] The step of identifying the set of interior points using the initial mesh of the anatomical surface can include: generating a plurality of points interior to a bounding box for the initial mesh; identifying a subset of the plurality of points that are interior to both the initial meshAttorney Docket No. 16043 WOO 1 / 82410- 1319and a convex hull of the set of surface points; and defining the subset of the plurality of points as the set of interior points. The plurality of points interior to the bounding box for the initial mesh may be randomly distributed within the bounding box for the initial mesh.
[0013] The step of constructing the model of the anatomical surface from the set of surface points and the set of interior points can include applying an alpha shape algorithm to the set of surface points and the set of interior points.
[0014] In other aspects, the step of constructing the initial mesh of the anatomical surface from the set of surface points can include the electroanatomical mapping system applying a patch geometry algorithm to the set of surface points.
[0015] Also disclosed herein is an electroanatomical mapping system including a display; and a surface modeling module configured to: receive a plurality of ultrasound images of an anatomical volume; generate a set of surface points from the plurality of ultrasound images; construct an initial mesh of an anatomical surface of the anatomical volume from the set of surface points; identify a set of interior points using the initial mesh of the anatomical surface; construct a model of the anatomical surface from the set of surface points and the set of interior points; and output a graphical representation of the model of the anatomical surface to the display.
[0016] The plurality of ultrasound images may include a plurality of two-dimensional ultrasound images.
[0017] The surface modeling module may be configured to generate the set of surface points from the plurality of ultrasound images by executing a process including: receiving a user input identifying a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; and discretizing the identified blood pool / tissue boundary of the anatomical volume into the set of surface points. Alternatively or additionally, the surface modeling module may be configured to generate the set of surface points from the plurality of ultrasound images by executing a process including: segmenting a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; and discretizing the segmented blood pool / tissue boundary of the anatomical volume into the set of surface points.
[0018] The surface modeling module may be configured to construct the initial mesh of an anatomical surface of the anatomical volume from the set of surface points by executing aAttorney Docket No. 16043 WOO 1 / 82410- 1319process including: estimating normals for the set of surface points; and generating the initial mesh of the anatomical surface using a Poisson surface reconstruction algorithm. The process may further include cropping the initial mesh of the anatomical surface to a bounding box for the set of surface points to generate a cropped initial mesh. Still further, the process may include applying a hole-filling algorithm to the initial mesh of the anatomical surface to generate a filled initial mesh. In yet further aspects, the process can include resampling the set of surface points across the initial mesh of the anatomical surface, such as resampling the set of surface points uniformly across the initial mesh of the anatomical surface.
[0019] The surface modeling module may be configured to identify the set of interior points using the initial mesh of the anatomical surface by executing a process including: generating a plurality of points interior to a bounding box for the initial mesh; identifying a subset of the plurality of points that are interior to both the initial mesh and a convex hull of the set of surface points; and defining the subset of the plurality of points as the set of interior points. The plurality of points interior to the bounding box for the initial mesh may be randomly distributed within the bounding box for the initial mesh.
[0020] The modeling module may be configured to construct the model of the anatomical surface from the set of surface points and the set of interior points by applying an alpha shape algorithm to the set of surface points and the set of interior points.
[0021] In other aspects, the modeling module may be configured to construct the initial mesh of the anatomical surface from the set of surface points by applying a patch geometry algorithm to the set of surface points.
[0022] The instant disclosure also provides a computer-implemented method of modeling an anatomical surface bounding an anatomical volume. The method includes receiving a plurality of ultrasound images of the anatomical volume in an electroanatomical mapping system, the el ectroan atomical mapping system including a display; via the electroanatomical mapping system: generating a set of surface points from the plurality of ultrasound images; constructing an initial mesh of the anatomical surface from the set of surface points; cropping the initial mesh of the anatomical surface to a bounding box for the set of surface points to create a cropped initial mesh; filling holes within the cropped initial mesh to create a filled initial mesh; identifying a set of interior points using the filled initial mesh; and constructing a model of theAttorney Docket No. 16043 WOO 1 / 82410- 1319anatomical surface from the set of surface points and the set of interior points; and outputting a graphical representation of the model of the anatomical surface to the display.
[0023] There is also provided a computer-readable medium, a record carrier, or a computer program product comprising instructions that, when executed, cause a computer or processor to perform any of the methods set forth herein. It will also be appreciated that the methods undertaken herein, including the segmentations and computations, may be undertaken by a processor or a computer on data representative of the received images.
[0024] The foregoing and other aspects, features, details, utilities, and advantages of the present invention will be apparent from reading the following description and claims, and from reviewing the accompanying drawings.BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 is a schematic diagram of an exemplary electroanatomical mapping system.
[0026] Figure 2 is a flowchart of representative steps that can be carried out according to aspects of the instant disclosure.
[0027] Figures 3A and 3B are different two-dimensional ultrasound views of a left ventricle including contour lines denoting the blood pool / tissue boundary.
[0028] Figure 4A illustrates the discretization of the contour line of Figure 3 A into a set of surface points.
[0029] Figure 4B positions the set of surface points of Figure 4A in a three-dimensional coordinate system.
[0030] Figure 4C shows a set of surface points discretized from a collection of six contour lines taken from respective ultrasound images.
[0031] Figure 5A shows an initial surface mesh generated by applying a Poisson surface reconstruction algorithm to the set of surface points of Figure 4C.
[0032] Figure 5B is a cropped initial surface mesh generated by cropping the initial surface mesh of Figure 5A to a bounding box for the set of surface points of Figure 4C.
[0033] Figure 5C illustrates a set of surface points uniformly resampled across the cropped initial surface mesh of Figure 5B.
[0034] Figure 5D illustrates a hole-filled initial surface mesh.Attorney Docket No. 16043 WOO 1 / 82410- 1319
[0035] Figure 6 illustrates the uniformly-resampled set of surface points of Figure 5C with a set of interior points.
[0036] Figure 7 illustrates a geometric surface model generated by applying an alpha-shapes algorithm to the combined set of points of Figure 6, with the initial contour lines superimposed thereon.
[0037] Figure 8A shows a set of surface points generated by discretizing a plurality of contour lines.
[0038] Figure 8B shows an initial surface mesh generated by applying a Poisson surface reconstruction algorithm to the set of surface points of Figure 8 A.
[0039] Figure 8C illustrates a set of surface points uniformly resampled across the initial surface mesh of Figure 8B.
[0040] Figure 8D illustrates the uniformly-resampled set of surface points of Figure 8C with a set of interior points.
[0041] Figure 8E illustrates a geometric surface model generated by applying an alphashapes algorithm to the combined set of points of Figure 8D, with the initial contour lines superimposed thereon.
[0042] Figure 9 is a flowchart of representative steps that can be carried out according to additional aspects of the instant disclosure that may be advantageous when modeling small and / or non-convex anatomical surfaces.
[0043] Figure 10 illustrates various papillary muscle geometries.
[0044] Figure 11 is an ultrasound image including a papillary muscle.
[0045] Figures 12A through 12D illustrate an approach to splitting a non-convex M-shaped surface contour into constituent surface contours based on curvature.
[0046] Figures 13 A and 13B illustrate the creation of convex surface contours (Figure 13B) from a set of surface contours including both convex and non-convex surface contours (Figure 13A).
[0047] Figures 13C-1 through 13C-3 are various views of a surface model created from the surface contours of Figure 13B.
[0048] Figure 14A is an ultrasound image including a mitral valve.Attorney Docket No. 16043 WOO 1 / 82410- 1319
[0049] Figure 14B is a set of geometry points created by discretizing the surface contour of the mitral valve as shown in Figure 14 A.
[0050] Figure 15 illustrates the application of a narrow-band surface reconstruction algorithm as disclosed herein to the discretized mitral valve surface contour of Figure 14B.
[0051] Figure 16 depicts an anatomical surface model created from the set of surface contours of Figure 15.
[0052] Figure 17 is a representative set of discretized surface contours for a left atrial appendage.
[0053] Figure 18 depicts an anatomical surface model created from the set of surface contours of Figure 17.
[0054] While multiple embodiments are disclosed, still other embodiments of the present disclosure will become apparent to those skilled in the art from the following detailed description, which shows and describes illustrative embodiments. Accordingly, the drawings and detailed description are to be regarded as illustrative in nature and not restrictive.DETAILED DESCRIPTION
[0055] The instant disclosure provides systems, apparatuses, and methods for modeling anatomical surfaces bounding anatomical volumes, such as endocardial surfaces bounding heart chambers. Reference will be made herein to the use of two-dimensional intracardiac echocardiography (ICE) images. Two-dimensional ICE volumetric images may be collected using an ICE catheter, such as Abbott Laboratories’ ViewFlex™ Xtra ICE catheter (Abbott Park, Illinois). The disclosure herein is provided in the context of a procedure carried out using an electroanatomical mapping system, such as the EnSite Precision™ cardiac mapping system or the Ensite™ X EP System, both also from Abbott Laboratories.
[0056] Those of ordinary skill in the art will understand, however, how to apply the teachings herein to good advantage in other contexts and / or with respect to other devices. For instance, the ordinarily-skilled artisan will appreciate that the teachings herein can be applied to other types of two-dimensional ultrasound images, including, without limitation, transesophageal echocardiographic (TEE) images and transthoracic echocardiographic (TTE) volumetric images. Likewise, the ordinarily-skilled artisan will appreciate how to extend the teachings herein toAttorney Docket No. 16043 WOO 1 / 82410- 1319three-dimensional volumetric images as well as to two- and three-dimensional images of anatomical regions other than the heart.
[0057] Figure 1 shows a schematic diagram of an exemplary electroanatomical mapping system 8 for conducting cardiac electrophysiology procedures, such as electrophysiological mapping and ablation. System 8 can be used, for example, to create an anatomical model of the patient’s heart 10 using one or more electrodes. System 8 can also be used to measure electrophysiology data at a plurality of points along a cardiac surface and store the measured data in association with location information for each measurement point at which the electrophysiology data was measured, for example to create a diagnostic data map of the patient’s heart 10.
[0058] As one of ordinary skill in the art will recognize, system 8 determines the location, and in some aspects the orientation, of objects, typically within a three-dimensional space, and expresses those locations as position information determined relative to at least one reference. This is referred to herein as “localization.”
[0059] As depicted in Figure 1 and described herein, system 8 can be a hybrid system that incorporates both impedance-based and magnetic field-based localization capabilities. The system 8 may be the EnSite™ Velocity™ or EnSite Precision™ cardiac mapping system or the Ensite™ XEP System, all from Abbott Laboratories. Other electroanatomical mapping systems, however, may be used in connection with the present teachings, including, for example, the Mediguide Ltd. gMPS system (also of Abbott Laboratories), the RHYTHMIA HDX™ and OPAL HDx™ mapping systems of Boston Scientific Corporation (Marlborough, Massachusetts), the CARTO navigation and localization system of Biosense Webster, Inc.(Irvine, California), the AURORA® system of Northern Digital Inc. (Waterloo, Ontario, Canada), Stereotaxis, Inc.’s (St. Louis, Missouri) NIOBE® Magnetic Navigation System, the KODEX® system (Koninklijke Philips N.V.; Amsterdam, Netherlands), the Affera™ Mapping and Ablation System of Medtronic pic (Minneapolis, Minnesota), and the Globe® PF Mapping and Ablation System of Kardium Inc. (Burnaby, British Columbia, Canada).
[0060] The localization and mapping systems described in the following patents (all of which are hereby incorporated by reference in their entireties) can also be used with the instant teachings: United States Patent Nos. 6,990,370 (filed 12 April 2000); 6,978,168 (filed 26Attorney Docket No. 16043 WOO 1 / 82410- 1319February 2003); 6,947,785 (filed 12 April 2000); 6,939,309 (filed 12 April 2000); 6,728,562 (filed 12 April 2000); 6,640,119 (filed 12 April 2000); 5,983,126 (filed 1 August 1997); and 5,697,377 (filed 22 November 1995).
[0061] The foregoing systems, and the modalities they employ to localize a medical device, will be familiar to those of ordinary skill in the art. Insofar as the ordinarily-skilled artisan will appreciate the basic operation of such systems, therefore, they are only described herein to the extent necessary to understand the instant disclosure.
[0062] For simplicity of illustration, the patient 11 is depicted schematically as an oval. As shown in Figure 1, three sets of surface electrodes (e.g, patch electrodes) 12, 14, 16, 18, 19, and 22 are shown applied to a surface of the patient 11, pairwise defining three generally orthogonal axes, referred to herein as an x-axis (12, 14), a y-axis (18, 19), and a z-axis (16, 22).Alternatively, the electrodes could be positioned in other arrangements, for example multiple electrodes on a particular body surface. As a further alternative, the electrodes do not need to be on the body surface but could be positioned internally to the body. Regardless of configuration, the patient’s heart 10 lies within the electric field generated by patch electrodes 12, 14, 16, 18, 19, and 22.
[0063] Figure 1 also depicts a magnetic source 30, which is coupled to magnetic field generators. In the interest of clarity, only two magnetic field generators 32 and 33 are depicted in Figure 1, but additional magnetic field generators (e.g., a total of six magnetic field generators, defining three generally orthogonal axes analogous to those defined by patch electrodes 12, 14, 16, 18, 19, and 22) can be used without departing from the scope of the present teachings.
[0064] An additional surface reference electrode (e.g, a “belly patch”) 21 provides a reference and / or ground electrode for the system 8. The belly patch electrode 21 may be an alternative to a fixed intra-cardiac electrode 31, described in further detail below. A magnetic patient reference sensor - anterior (“PRS-A”) can also be positioned on the patient’s chest to serve as a reference, analogous to surface reference electrode 21 and / or intracardiac reference electrode 31, for magnetic field-based localization modalities.
[0065] It should also be appreciated that, in addition, the patient 11 may have most or all of the conventional electrocardiogram (“ECG” or “EKG”) system leads in place. For example, a standard set of 12 ECG leads may be utilized for sensing electrocardiograms on the patient’sAttorney Docket No. 16043 WOO 1 / 82410- 1319heart 10. This ECG information is available to the system 8 (e.g., it can be provided as input to computer system 20). Insofar as ECG leads are well understood, and for the sake of clarity in the figures, only a single lead 6 and its connection to computer 20 is illustrated in Figure 1.
[0066] Representative catheters 13, 40 are also shown schematically in Figure 1. In aspects of the disclosure, catheter 13 can be an ablation catheter, such as the Abbott Laboratories FlexAbility™ Ablation Catheter, Sensor Enabled™, and catheter 40 can be an intracardiac echocardiography (ICE) catheter, such as the Abbott Laboratories ViewFlex™ Xtra ICE catheter. Catheters 13, 40 each respectively include one or more sensors 17, 42 for sensing the electric fields generated by patch electrodes 12, 14, 16, 18, 19, and 22 and / or the magnetic fields generated by magnetic field generators 32, 33.
[0067] An optional fixed reference electrode 31 e.g., attached to a wall of the heart 10) is shown on yet another catheter 29. Often, reference electrode 31 is placed in the coronary sinus and defines the origin of a coordinate system with reference to which catheters 13, 40 can be localized by system 8.
[0068] The computer 20 may comprise, for example, a conventional general-purpose computer, a special-purpose computer, a distributed computer, or any other type of computer. The computer 20 may comprise one or more processors 28, such as a single central processing unit (“CPU”), or a plurality of processing units, commonly referred to as a parallel processing environment, which may execute instructions to practice the various aspects described herein.
[0069] Amongst other things, computer 20 can interpret measurements by sensors 17, 42 of the magnetic and / or electrical fields generated by magnetic field generators 32, 33 and patch electrodes 12, 14, 16, 18, 19, and 22, respectively, to localize catheters 13, 40 within heart 10.
[0070] Ultrasound imaging catheter 40 can be used to collect a plurality of two-dimensional images of heart 10 using any of several echographic imaging modalities, such as B-mode ultrasound and color Doppler echocardiography. These two-dimensional images can optionally be assembled into a three-dimensional volumetric image of heart 10 (or other anatomic structure) using various techniques, including those disclosed in United States patent application publication no. 2006 / 0241445 (filed 26 April 2005) (which is hereby incorporated by reference as though fully set forth herein).Attorney Docket No. 16043 WOO 1 / 82410- 1319
[0071] It is contemplated that ultrasound imaging catheter 40 may be coupled to an ultrasound console, such as Abbott Laboratories’ ViewMate™ Multi Ultrasound Console, which may in turn be coupled to system 8. Alternatively, and for purposes of the disclosure herein, ultrasound imaging catheter 40 will be described as coupled directly to system 8, such that aspects of the disclosure can be carried out on processor(s) 28 of computer 20.
[0072] The foregoing discussion of ICE imaging is general, insofar as numerous aspects of ICE imaging, including the use of ICE imaging in connection with electrophysiology procedures, are well-understood by those of ordinary skill in the art and need not be described in detail herein. See, e.g., Enriquez et al., “Use of Intracardiac Echocardiography in Interventional Cardiology,” Circulation, Vol. 137, Issue 21, pp. 2278-2294 (May 22, 2018). Thus, ICE imaging will only be described herein to the extent necessary to understand the instant disclosure.
[0073] As mentioned above, aspects of the disclosure relate to generating anatomical surface models from two-dimensional medical images. System 8 can therefore include a surface modeling module 58, which may be software based (e.g., a series of programming instructions executed on processor(s) 28 of computer 20), hardware-based e.g., an application specific integrated circuit (ASIC)), or a combination thereof.
[0074] One exemplary method according to aspects of the instant disclosure will be explained with reference to the flowchart 200 of representative steps presented as Figure 2. For example, flowchart 200 may represent several exemplary steps that can be carried out by electroanatomical mapping system 8 of Figure 1 (e.g., by processor(s) 28 and / or surface modeling module 58). It should be understood that the representative steps described below can be either hardware- or software-implemented. For the sake of explanation, the term “signal processor” is used herein to describe both hardware- and software-based implementations of the teachings herein.
[0075] In block 202, system 8 receives a plurality of ultrasound images (e.g., two-dimensional ICE images) of the anatomical volume (e.g., a heart chamber) for which a surface model (e.g., a model of the endocardial surface) is desired. As mentioned briefly above, those of ordinary skill in the art will be familiar with various ultrasound imaging techniques, and any such technique is regarded as within the scope of block 202. In any case, Figures 3 A and 3B illustrate representative two-dimensional ICE images 300a, 300b respectively, of two differentAttorney Docket No. 16043 WOO 1 / 82410- 1319views of a left ventricle. It should be understood, however, that the left ventricle is chosen of illustrative purposes only and that the techniques disclosed herein can be applied to other cardiac chambers.
[0076] In block 204, system 8 generates a set of surface points from the ultrasound images received in block 202. As used herein, the term “surface points” refers to points that are on or very close to (e.g., within about 1 mm) the anatomical surface being modeled (e.g, the endocardial surface).
[0077] In certain aspects of the disclosure, system 8 can generate the set of surface points starting from a user input that identifies a blood pool / tissue boundary of the anatomical volume in the ultrasound images. The user input can be provided through a graphical user interface (GUI) that system 8 generates (e.g, on display 23).
[0078] For instance, system 8 can display the ultrasound images received in block 202 (e.g., two-dimensional ICE images 300a, 300b of Figures 3A and 3B) as part of a GUI that allows the user to draw (e.g, via a suitable input device, such as a mouse, connected to computer 20) a plurality of contour lines (302a, 302b in Figures 3 A and 3B, respectively) corresponding to the blood pool / tissue boundary of the anatomical volume on the ultrasound images themselves.
[0079] The user may, for example, draw at least two contour lines, at least four contour lines, at least six contour lines, or at least eight contour lines in various ultrasound images (e.g., different views of the left ventricle, such as Figures 3A and 3B).
[0080] Once the user has identified the contour lines, system 8 can discretize the contour lines into the set of surface points. For example, system 8 can generate the set of surface points by sampling the user-identified contour lines at an interval of between about 2 pixels and about 20 pixels, between about 2 pixels and about 10 pixels, or between about 2 pixels and about 5 pixels. As those of ordinary skill in the art will appreciate, coordinate values for the surface points discretized from the contour lines can be determined from the localization information for catheter 40 that is associated with the underlying ultrasound images.
[0081] Figure 4A illustrates the discretization of contour line 302a in two-dimensional ultrasound image 300a of Figure 3 A into a set of surface points 400. Figure 4B illustrates the set of surface points 400 plotted in a three-dimensional coordinate system 402 using the localization information associated with two-dimensional ultrasound image 300a in Figure 3A. Finally,Attorney Docket No. 16043 WOO 1 / 82410- 1319Figure 4C illustrates an expanded set of surface points 404 discretized from a collection of six contour lines taken from six different ultrasound images of the left ventricle, again plotted in a three-dimensional coordinate system 402 using the localization system associated with the corresponding ultrasound images. It should be noted that this initial input set of surface points 400 is typically relatively sparse in three-dimensional space.
[0082] In other aspects of the disclosure, system 8 can generate the set of surface points by first automatically identifying the blood pool / tissue boundary of the anatomical volume in the ultrasound images using segmentation techniques that will be familiar to those of ordinary skill in the art. Once the blood pool / tissue boundary has been identified, a discretization process similar or identical to that described above may be applied to generate the set of surface points from the automatically-segmented blood pool / tissue boundary.
[0083] Depending on the number of surface points in the set of surface points, system 8 may optionally downsample the set of surface points prior to further processing. For instance, the set of surface points may be placed into cubic voxels of side length h, where h may have a value of about 1.2 units (with “units” referring to the unit of measurement of the coordinate system employed of system 8, such as mm). System 8 can then resolve any surface points within each voxel into a single surface point, such as the centroid of all surface points within that voxel.
[0084] In block 206, system 8 uses the set of surface points to construct an initial mesh of the anatomical surface (e.g., the endocardial surface). According to aspects of the instant disclosure, system 8 uses a Poisson surface reconstruction algorithm to construct the initial mesh of the anatomical surface. The use of a Poisson equation to solve a three-dimensional shape reconstruction problem is described by Michael Kazhdan, Matthew Bolitho, and Hugues Hope in Poisson Stirface Reconstruction, Eurographics Symposium on Geometry Processing (2006), which is incorporated by reference as though fully set forth herein.
[0085] Those of ordinary skill in the art will understand that Poisson surface reconstruction algorithms utilize both surface points and their normals (that is, unit vectors normal to the surface being reconstructed) as inputs. The set of surface points generated in block 204, however, will not typically include normals. Thus, system 8 can estimate normals for the set of surface points, such as by using a KD-tree search, where the search parameters are a radius R ofAttorney Docket No. 16043 WOO 1 / 82410- 1319about between about 75 and about 125, such as about 100, and a maximum of nearest neighbors of between about 50 and about 60, such as about 55.
[0086] The normals estimated by the KD-tree search may also be oriented to ensure consistency across the set of surface points, such as by using the tangent plane method with a specified parameter k. k is known as the “Ann graph parameter” and refers to a number (k) of nearest neighbors used to build the Riemannian nearest neighbors graph that is, in turn, used to propagate normal orientation, k may be between about 24 and about 30, such as about 25.Further, the Riemannian nearest neighbors graph may be built using the set of surface points 400.
[0087] Application of the Poisson surface reconstruction algorithm in block 206 will output a smooth, closed, triangular surface mesh that approximates the set of surface points. A representative surface mesh 500 is shown in Figure 5A (with the set of surface points 404 shown superimposed thereon).
[0088] As those familiar with Poisson surface reconstructions will appreciate, the initial mesh may include certain artifacts (e.g., holes) for which post-processing in block 208 is desirable. Various post-processing functions, which may be executed individually or in combinations of two or more individual functions, are contemplated.
[0089] One contemplated post-processing function is a cropping function. In a cropping function, the initial mesh output from block 206 is cropped to a bounding box for the set of surface points. This generates a cropped initial mesh that omits extraneous parts of the initial mesh that lie outside the bounding box for the set of surface points. Figure 5B shows a cropped initial mesh 502 generated by applying a cropping function to the representative surface mesh 500 of Figure 5 A.
[0090] Another contemplated post-processing function is a resampling function. In a resampling function, the set of surface points is resampled (typically, but not necessarily, uniformly) across the surface of the initial mesh (or, where a cropping function is used, the surface of the cropped initial mesh). The number of resampled surface points may be user-selected, with a default value of about 20,000. The resampled surface points may further be limited to those that also lie within a convex hull of the original surface point set (that is, the point set generated in block 204). Figure 5C illustrates a set of surface points 504 uniformly resampled over the cropped initial mesh 502.Attorney Docket No. 16043 WOO 1 / 82410- 1319
[0091] Yet another contemplated post-processing function is a hole-filling function. In a hole-filling function, system 8 executes one or more iterations of a hole-filling algorithm on the initial mesh (or the cropped and / or resampled initial mesh) to fill holes therein and generate a filled initial mesh. Those of ordinary skill in the art will be familiar with various hole-filling algorithms. For purposes of illustration only, however, the Open3D open-source library includes various functions, such as a dilation function, that can be used to fill holes in a three-dimensional mesh. Figure 5D illustrates a hole-filled, cropped initial mesh 506.
[0092] In block 210, system 8 uses the initial mesh from block 206, subject to any postprocessing functions that may have been applied in block 208, to generate a set of interior points. As used herein, the term “interior points” refers to points that are inside the anatomical surface being modeled (e.g., points within the heart chamber).
[0093] Although those of ordinary skill in the art will appreciate that there are numerous suitable ways to generate a set of interior points, an illustrative approach is as follows. First, system 8 establishes a raycasting scene and adds the (post-processed) initial mesh to this scene. System 8 then generates random points within the bounding box of the (post-processed) initial mesh and, for each of these randomly-generated points, calculates the signed distance to the (post-processed) initial mesh. Points with a negative signed distance are deemed to be inside the (post-processed) initial mesh.
[0094] Next, the convex hull of the original surface point set (that is, the surface point set generated in block 204) is computed and a second raycasting scene is established for this convex hull. Points that are inside both the (post-processed) initial mesh (z.e., those with a negative signed distance) and the convex hull of the original surface point set are defined as the set of interior points. Figure 6 illustrates both the set of surface points 504 that is uniformly redistributed over the cropped initial mesh and the set of interior points 600.
[0095] In block 212, system 8 generates a model of the anatomical surface (e.g, the endocardial surface) from the union of the set of surface points (e.g., after post-processing in block 208) and the set of interior points (e.g., as output from block 210). There are various known algorithms and techniques for generating a surface from a point cloud, all of which are regarded as within the scope of the instant disclosure. For purposes of illustration, however, it isAttorney Docket No. 16043 WOO 1 / 82410- 1319contemplated that an alpha-shape algorithm may be applied to the union of the set of surface points and the set of interior points to generate the model of the anatomical surface.
[0096] System 8 may output a graphical representation of the model of the anatomical surface (e.g., on display 23) in block 214. Figure 7 is an illustrative graphical representation 700 of the model of the anatomical surface. To illustrate the accuracy of the model, the original contour lines 702 are shown superimposed on graphical representation 700.
[0097] Figures 8A-8E provide an additional illustration of the process described above. Namely:• Figure 8A illustrates a set of surface points 800 taken from a plurality of contours; • Figure 8B illustrates an initial mesh of a surface 802 generated by applying a Poisson surface reconstruction algorithm to the set of surface points 800 (surface points 800 are also shown in Figure 8B);• Figure 8C shows a set of surface points 804 uniformly resampled over the Poisson surface 802;• Figure 8D additionally shows a set of interior points 806;• Figure 8E shows a model 808 of a surface generated by applying an alpha-shape algorithm to the set of surface points 804 and the set of interior points 806 with the initial surface points 800 superimposed thereon.
[0098] The approach described above will create a surface, enclosing a solid volume, with a consistent geometric shape for most surface point sets. Certain edge cases, however, merit additional discussion herein. For instance, if the initial surface point set (that is, the point set generated in block 204) is substantially coplanar (because, for example, the user specified only a single contour line in the ultrasound images), the construction of the initial mesh in block 206 will differ from the foregoing description, which references a Poisson surface reconstruction algorithm.
[0099] In particular, if the initial surface point set is substantially coplanar, system 8 may instead apply a patch geometry algorithm to construct the initial mesh in block 206. Patch geometry algorithms are described, for example, in United States patent nos. 9,947,142 (filed 20Attorney Docket No. 16043 WOO 1 / 82410- 1319October 2015) and 10,861,246 (filed 8 March 2018), both of which are incorporated by reference as though fully set forth herein.
[0100] Variations of the teachings above may be applied to good advantage to model small and / or non-convex anatomical surfaces. Examples of such surfaces include, without limitation, papillary muscles, the left atrial appendage (LAA), cardiac valves (e.g., mitral valve, tricuspid valve, pulmonary valves, aortic valves), and pulmonary veins (PVs).
[0101] Through localization of ultrasound imaging catheter 40, system 8 (e.g., surface modeling module 58 introduced above) may be able to determine a priori whether certain received ultrasound images are of such small and / or non-convex geometries and automatically apply a suitable variation as described below (e.g., if the localization of ultrasound imaging catheter 40 indicates that the ultrasound images are of papillary muscles, system 8 can apply the variation for papillary muscle surface modeling described below). Of course, it is also contemplated that the practitioner may select the particular surface modeling technique to apply to a set of received ultrasound images from the variations disclosed herein.
[0102] In any event, a general overview of the creation of models of small and / or non-convex anatomical surfaces will be provided with reference to flowchart 900 of Figure 9, which may represent several exemplary steps that can be carried out by electroanatomical mapping system 8 of Figure 1 (e.g., by processor(s) 28 and / or surface modeling module 58). It should be understood that the representative steps described below can be either hardware- or software-implemented. For the sake of explanation, the term “signal processor” is used herein to describe both hardware- and software-based implementations of the teachings herein.
[0103] Block 202, in which system 8 receives a plurality of ultrasound images, is described above. In block 902, system 8 (e.g., surface modeling module 58) generates a plurality of contours of the anatomical surface from the plurality of ultrasound images received in block 202. Various approaches to generating such surface contours are also described above.
[0104] In block 904, system 8 (e.g., surface modeling module 58) classifies the surface contours into two (or more) subsets. For convenience, the two subsets will be referred to herein as a first subset and a second subset; one of ordinary skill in the art will appreciate how to extend the teachings herein to additional subsets of surface contours. In some aspects of the disclosure,Attorney Docket No. 16043 WOO 1 / 82410- 1319the first subset may include convex surface contours and the second subset may include non-convex surface contours.
[0105] In blocks 906a, 906b, system 8 (e.g., surface modeling module 58) generates first and second sub-models of the anatomical surface from the first and second subsets of the surface contours, respectively. Two different surface modeling algorithms are applied, however, to generate the first and second sub-models of the anatomical surface. For example, the surface modeling algorithm described above may be applied to generate the first sub-model (e.g., of the convex surface contours), while one of the surface modeling algorithms described below may be applied to generate the second sub-model (e.g., of the non-convex surface contours).
[0106] System 8 e.g., surface modeling module 58) fuses the first and second sub-models in block 908. The fused model may be output (e.g., to display 23) in block 214 as described above.
[0107] Surface Modeling for Papillary Muscles
[0108] Figure 10 illustrates that papillary muscles geometries can be classified into various shapes, including convex non-M-shaped 1000, non-convex M-shaped 1002, and non-convex multi-peak 1004 (which, for purposes of the instant disclosure, can be regarded as multiple M-shaped). Thus, when modeling the surface of a papillary muscle, block 904 may include classifying the ultrasound images into a first subset of convex non-M-shaped contours and a second subset of non-convex M-shaped contours (including non-convex multi-peak shapes).
[0109] As mentioned above, the surface modeling algorithm described above may be applied to the first subset of convex non-M-shaped contours. On the other hand, the surface algorithm applied to the second subset of non-convex M-shaped contours may include splitting each non-convex M-shaped contour into two (or more) convex non-M-shaped contours to which the surface modeling algorithm described above may ultimately be applied.
[0110] For instance, the non-convex M-shaped contours may be discretized into a plurality of geometry points, analogous to the discussion of block 204 above. System 8 (e.g., surface modeling module 58) may then detect one or more high-curvature geometry points and split the respective non-convex M-shaped contour into two (or more) constituent convex non-M-shaped contours at the detected high-curvature geometry points.Attorney Docket No. 16043 WOO 1 / 82410- 1319
[0111] Figure 11 depicts an ultrasound image 1100 of a papillary muscle and the non-convex M-shaped contour 1102 thereof. Figure 12A is a representative discretization 1200 of non-convex M-shaped contour 1102 into a plurality of geometry points 1202. Figure 12B identifies several high-curvature geometry points 1204 (shown solid, as opposed to the outline-only representation of other geometry points) amongst the plurality of geometry points 1202. In Figure 12C, the three highest-curvature geometry points 1206 - representing the valley and two surrounding peaks of the M-shape - are identified. In Figure 12D, non-convex M-shaped contour 1102 is split at the high-curvature geometry point 1208 in the valley of the M-shape into con stituent contours 1210a, 1210b.
[0112] There are various approaches to detecting the high-curvature geometry points that may be suitable. For example, the curvature at each geometry point may be computed using finite difference methods.
[0113] In some embodiments of the disclosure, however, system 8 (e.g., surface modeling module 58) may determine the curvature at each geometry point, and thereby detect the high-curvature geometry points, as a function of an osculating circle radius at each geometry point. The use of osculating circle radius to detect high-curvature geometry points is advantageous by comparison to other techniques (e.g., finite difference methods) because it offers a very robust geometric measure of curvature.
[0114] The osculating circle at each geometry point may be defined as follows:• For each geometry point Pt, identify the preceding geometry point Pt i moving backward along non-convex M-shaped contour 1102 and the succeeding geometry point Pi+i moving forward along non-convex M-shaped contour 1102.• Define the triangle with vertices Pi, Pi-i, and Pt-i. The osculating circle is the d^b *ccircumcircle of this triangle and has radius r = , where a, b, and c are the lengthsof the sides of the triangle and A is the area of the triangle. (The radius may also be comp1uted as r = — - — .)2*sin>l7
[0115] Once the non-convex M-shaped contours have been split into constituent convex non-M-shaped contours, system 8 (e.g., surface modeling module 58) may generate a surfaceAttorney Docket No. 16043 WOO 1 / 82410- 1319model therefrom according to the teachings above (e.g., using a Poisson surface reconstruction algorithm to construct an initial mesh and so forth).
[0116] Figures 13A, 13B, and 13C-1 through 13C-3 illustrate surface model construction for an anatomy including a papillary muscle according to the foregoing teachings. Figure 13 A illustrates surface contours 1300, including both convex non-M-shaped contours 1302 and a non-convex M-shaped contour 1304. Figure 13B illustrates non-convex M-shaped contour 1304 once split into constituent convex contours 1304a, 1304b according to the foregoing teachings. Figures 13C-1 through 13C-3 are various views of a surface model 1306 generated from the surface contours of Figure 13B (the source surface contours are also shown in Figures 13C-1 through 13C-3).
[0117] Surface Modeling for Cardiac Valves
[0118] In another aspect, the instant disclosure provides techniques for modeling very thin non-convex anatomical surfaces, such as cardiac valves, that may have only a single surface contour. Figure 14A is an ultrasound image 1400 including a view of the mitral valve with the surface contour 1402 of the mitral valve annotated thereon. Figure 14B shows surface contour 1402 discretized into a plurality of geometry points 1404.
[0119] As previously discussed, the surface modeling algorithm described above may be applied to the first subset of convex contours (e.g., other than the non-convex contour for the cardiac valve). On the other hand, the surface modeling algorithm applied to the second subset of non-convex contour of the cardiac may include a narrow-band surface reconstruction algorithm.
[0120] In an exemplary narrow-band surface reconstruction algorithm, system 8 (e.g., surface modeling module 58) generates two additional surface contours from the single surface contour 1402 of the cardiac valve by first estimating a normal to each geometry point 1402 (e.g., using a KD-tree search as described above) and then offsetting each geometry point by a small distance (e.g., between about 2mm and about 3mm) in both the positive- and negative-normal directions to create additional discretized surface contours 1500a, 1500b as shown in Figure 15.
[0121] Once the original non-convex contour has been so replicated, system 8 (e.g., surface modeling module 58) may generate a surface model from the set of three contours according toAttorney Docket No. 16043 WOO 1 / 82410- 1319the teachings above (e.g., using a Poisson surface reconstruction algorithm to construct an initial mesh and so forth), such as surface model 1600 of Figure 16.
[0122] Surface Modeling for LAA and / or PVs
[0123] In yet another aspect, the instant disclosure provides techniques for modeling anatomical surfaces that are small relative to the cardiac chambers, such as the LAA (which may be treated as non-convex) or the PVs (which may be treated as convex). For such geometries it may be advantageous to apply the surface modeling algorithm described above while employing a smaller neighborhood for normal estimation (e.g., using the KD-tree search) than described above for larger and / or convex geometries.
[0124] For instance, the KD-tree search parameters described above are a radius R of between about 75 and about 125, such as about 100, and a maximum of nearest neighbors of between about 50 and about 60, such as about 55. For a relatively small and / or non-convex geometry, however, it may be desirable to use a radius R of between about 5 and 15, such as about 10, and a maximum of nearest neighbors of between about 10 and about 30, such as about 20.
[0125] Figure 17 depicts surface contours 1700 for a representative LAA. Figure 18 is a surface model 1800 generated therefrom according to the foregoing description (surface contours 1700 are also shown in Figure 18).
[0126] For completeness, the methods described herein may be methods that are embedded within a set of human- or machine-readable instructions that are comprised within a computer-readable medium or record carrier, or that are comprised within a computer program product. The instructions are such that, when executed by a computer or processor, such as a processor within the electroanatomical mapping system as described herein or a processor of a general-purpose computer system, the computer or processor causes the system to perform the methods described herein.
[0127] Although several embodiments have been described above with a certain degree of particularity, those skilled in the art could make numerous alterations to the disclosed embodiments without departing from the spirit or scope of this invention.
[0128] All directional references (e.g., upper, lower, upward, downward, left, right, leftward, rightward, top, bottom, above, below, vertical, horizontal, clockwise, andAttorney Docket No. 16043 WOO 1 / 82410- 1319counterclockwise) are only used for identification purposes to aid the reader’s understanding of the present invention, and do not create limitations, particularly as to the position, orientation, or use of the invention. Joinder references (e.g., attached, coupled, connected, and the like) are to be construed broadly and may include intermediate members between a connection of elements and relative movement between elements. As such, joinder references do not necessarily infer that two elements are directly connected and in fixed relation to each other.
[0129] It is intended that all matter contained in the above description or shown in the accompanying drawings shall be interpreted as illustrative only and not limiting. Changes in detail or structure may be made without departing from the spirit of the invention as defined in the appended claims.
[0130] The invention is defined in the appended claims. A non-exhaustive list of aspects of the invention set out in the numbered clauses is useful for understanding the invention:
[0131] Clause 1 : A computer-implemented method of modeling an anatomical surface bounding an anatomical volume, the method comprising: receiving a plurality of ultrasound images of the anatomical volume in an electroanatomical mapping system, the electroanatomical mapping system including a display; via the electroanatomical mapping system: generating a set of surface points from the plurality of ultrasound images; constructing an initial mesh of the anatomical surface from the set of surface points; identifying a set of interior points using the initial mesh of the anatomical surface; and constructing a model of the anatomical surface from the set of surface points and the set of interior points; and outputting a graphical representation of the model of the anatomical surface to the display.
[0132] Clause 2: The method of clause 1, wherein the plurality of ultrasound images comprises a plurality of two-dimensional ultrasound images.
[0133] Clause 3: The method of any preceding clause, wherein generating the set of surface points comprises the electroanatomical mapping system:receiving a user input identifying a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; andAttorney Docket No. 16043 WOO 1 / 82410- 1319discretizing the identified blood pool / tissue boundary of the anatomical volume into the set of surface points.
[0134] Clause 4: The method of any preceding clause, wherein generating the set of surface points comprises the electroanatomical mapping system:segmenting a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; anddiscretizing the segmented blood pool / tissue boundary of the anatomical volume into the set of surface points.
[0135] Clause 5: The method of any preceding clause, wherein constructing the initial mesh of the anatomical surface from the set of surface points comprises the electroanatomical mapping system:estimating normals for the set of surface points; andgenerating the initial mesh of the anatomical surface using a Poisson surface reconstruction algorithm.
[0136] Clause 6: The method of clause 5, wherein constructing the initial mesh of the anatomical surface from the set of surface points further comprises cropping the initial mesh of the anatomical surface to a bounding box for the set of surface points to generate a cropped initial mesh.
[0137] Clause 7: The method of either clause 5 or 6, wherein constructing the initial mesh of the anatomical surface from the set of surface points further comprises applying a hole-filling algorithm to the initial mesh of the anatomical surface to generate a filled initial mesh.
[0138] Clause 8: The method of any of clauses 5 to 7, further comprising resampling the set of surface points across the initial mesh of the anatomical surface.Attorney Docket No. 16043 WOO 1 / 82410- 1319
[0139] Clause 9: The method of clause 8, wherein resampling the set of surface points across the initial mesh of the anatomical surface comprises resampling the set of surface points uniformly across the initial mesh of the anatomical surface.
[0140] Clause 10: The method of any preceding clause, wherein identifying the set of interior points using the initial mesh of the anatomical surface comprises:generating a plurality of points interior to a bounding box for the initial mesh; identifying a subset of the plurality of points that are interior to both the initial mesh and a convex hull of the set of surface points; anddefining the subset of the plurality of points as the set of interior points.
[0141] Clause 11 : The method of clause 10, wherein the plurality of points interior to the bounding box for the initial mesh are randomly distributed within the bounding box for the initial mesh.
[0142] Clause 12: The method of any preceding clause, wherein constructing the model of the anatomical surface from the set of surface points and the set of interior points comprises applying an alpha shape algorithm to the set of surface points and the set of interior points.
[0143] Clause 13: The method of any preceding clause, wherein constructing the initial mesh of the anatomical surface from the set of surface points comprises the electroanatomical mapping system applying a patch geometry algorithm to the set of surface points.
[0144] Clause 14: An electroanatomical mapping system, comprising:a display; anda surface modeling module configured to:receive a plurality of ultrasound images of an anatomical volume;generate a set of surface points from the plurality of ultrasound images;Attorney Docket No. 16043 WOO 1 / 82410- 1319construct an initial mesh of an anatomical surface of the anatomical volume from the set of surface points;identify a set of interior points using the initial mesh of the anatomical surface; construct a model of the anatomical surface from the set of surface points and the set of interior points; andoutput a graphical representation of the model of the anatomical surface to the display.
[0145] Clause 15: The electroanatomical mapping system of clause 14, wherein the plurality of ultrasound images comprises a plurality of two-dimensional ultrasound images.
[0146] Clause 16: The electroanatomical mapping system of either clause 14 or 15, wherein the surface modeling module is configured to generate the set of surface points from the plurality of ultrasound images by executing a process comprising:receiving a user input identifying a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; anddiscretizing the identified blood pool / tissue boundary of the anatomical volume into the set of surface points.
[0147] Clause 17: The electroanatomical mapping system of any of clauses 14 to 16, wherein the surface modeling module is configured to generate the set of surface points from the plurality of ultrasound images by executing a process comprising:segmenting a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; anddiscretizing the segmented blood pool / tissue boundary of the anatomical volume into the set of surface points.
[0148] Clause 18: The electroanatomical mapping system of any of clauses 14 to 17, wherein the surface modeling module is configured to construct the initial mesh of an anatomicalAttorney Docket No. 16043 WOO 1 / 82410- 1319surface of the anatomical volume from the set of surface points by executing a process comprising:estimating normals for the set of surface points; andgenerating the initial mesh of the anatomical surface using a Poisson surface reconstruction algorithm.
[0149] Clause 19: The electroanatomical mapping system of clause 18, wherein the surface modeling module is further configured to crop the initial mesh of the anatomical surface to a bounding box for the set of surface points to generate a cropped initial mesh.
[0150] Clause 20: The electroanatomical mapping system of either clause 18 or 19, wherein the surface modeling module is further configured to apply a hole-filling algorithm to the initial mesh of the anatomical surface to generate a filled initial mesh.
[0151] Clause 21: The electroanatomical mapping system of any of clauses 18 to 20, wherein the surface modeling module is further configured to resample the set of surface points across the initial mesh of the anatomical surface.
[0152] Clause 22: The electroanatomical mapping system of clause 21, wherein the surface modeling module is configured to resample the set of surface points uniformly across the initial mesh of the anatomical surface.
[0153] Clause 23: The electroanatomical mapping system of any of clauses 14 to 22, wherein the surface modeling module is configured to identify the set of interior points using the initial mesh of the anatomical surface by executing a process comprising:generating a plurality of points interior to a bounding box for the initial mesh; identifying a subset of the plurality of points that are interior to both the initial mesh and a convex hull of the set of surface points; anddefining the subset of the plurality of points as the set of interior points.Attorney Docket No. 16043 WOO 1 / 82410- 1319
[0154] Clause 24: The electroanatomical mapping system of clause 23, wherein the plurality of points interior to the bounding box for the initial mesh are randomly distributed within the bounding box for the initial mesh.
[0155] Clause 25: The electroanatomical mapping system of any of clauses 14 to 24, wherein the modeling module is configured to construct the model of the anatomical surface from the set of surface points and the set of interior points by applying an alpha shape algorithm to the set of surface points and the set of interior points.
[0156] Clause 26: The electroanatomical mapping system according of any of clauses 14 to 25, wherein the modeling module is configured to construct the initial mesh of the anatomical surface from the set of surface points by applying a patch geometry algorithm to the set of surface points.
[0157] Clause 27: A computer-implemented method of modeling an anatomical surface bounding an anatomical volume, the method comprising:receiving a plurality of ultrasound images of the anatomical volume in an electroanatomical mapping system, the electroanatomical mapping system including a display;via the electroanatomical mapping system:generating a set of surface points from the plurality of ultrasound images; constructing an initial mesh of the anatomical surface from the set of surface points;cropping the initial mesh of the anatomical surface to a bounding box for the set of surface points to create a cropped initial mesh;filling holes within the cropped initial mesh to create a filled initial mesh; identifying a set of interior points using the filled initial mesh; and constructing a model of the anatomical surface from the set of surface points and the set of interior points; andAttorney Docket No. 16043 WOO 1 / 82410- 1319outputting a graphical representation of the model of the anatomical surface to the display.
Claims
Attorney Docket No. 16043 WOO 1 / 82410- 1319CLAIMSWhat is claimed is:
1. A computer-implemented method of modeling an anatomical surface bounding an anatomical volume, the method comprising:receiving a plurality of ultrasound images of the anatomical volume in an electroanatomical mapping system, the electroanatomical mapping system including a display;via the electroanatomical mapping system:generating a set of surface points from the plurality of ultrasound images; constructing an initial mesh of the anatomical surface from the set of surface points;identifying a set of interior points using the initial mesh of the anatomical surface; andconstructing a model of the anatomical surface from the set of surface points and the set of interior points; andoutputting a graphical representation of the model of the anatomical surface to the display.
2. The method according to claim 1, wherein the plurality of ultrasound images comprises a plurality of two-dimensional ultrasound images.
3. The method according to claim 1, wherein generating the set of surface points comprises the electroanatomical mapping system:receiving a user input identifying a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; anddiscretizing the identified blood pool / tissue boundary of the anatomical volume into the set of surface points.Attorney Docket No. 16043 WOO 1 / 82410- 13194. The method according to claim 1, wherein generating the set of surface points comprises the electroanatomical mapping system:segmenting a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; anddiscretizing the segmented blood pool / tissue boundary of the anatomical volume into the set of surface points.
5. The method according to claim 1, wherein constructing the initial mesh of the anatomical surface from the set of surface points comprises the electroanatomical mapping system:estimating normals for the set of surface points; andgenerating the initial mesh of the anatomical surface using a Poisson surface reconstruction algorithm.
6. The method according to claim 5, wherein constructing the initial mesh of the anatomical surface from the set of surface points further comprises cropping the initial mesh of the anatomical surface to a bounding box for the set of surface points to generate a cropped initial mesh.
7. The method according to claim 5, wherein constructing the initial mesh of the anatomical surface from the set of surface points further comprises applying a hole-filling algorithm to the initial mesh of the anatomical surface to generate a filled initial mesh.
8. The method according to claim 5, further comprising resampling the set of surface points across the initial mesh of the anatomical surface.
9. The method according to claim 8, wherein resampling the set of surface points across the initial mesh of the anatomical surface comprises resampling the set of surface points uniformly across the initial mesh of the anatomical surface.
10. The method according to claim 1, wherein identifying the set of interior points using the initial mesh of the anatomical surface comprises:generating a plurality of points interior to a bounding box for the initial mesh;Attorney Docket No. 16043 WOO 1 / 82410- 1319identifying a subset of the plurality of points that are interior to both the initial mesh and a convex hull of the set of surface points; anddefining the subset of the plurality of points as the set of interior points.
11. The method according to claim 10, wherein the plurality of points interior to the bounding box for the initial mesh are randomly distributed within the bounding box for the initial mesh.
12. The method according to claim 1, wherein constructing the model of the anatomical surface from the set of surface points and the set of interior points comprises applying an alpha shape algorithm to the set of surface points and the set of interior points.
13. The method according to claim 1, wherein constructing the initial mesh of the anatomical surface from the set of surface points comprises the electroanatomical mapping system applying a patch geometry algorithm to the set of surface points.
14. An electroanatomical mapping system, comprising:a display; anda surface modeling module configured to:receive a plurality of ultrasound images of an anatomical volume;generate a set of surface points from the plurality of ultrasound images; construct an initial mesh of an anatomical surface of the anatomical volume from the set of surface points;identify a set of interior points using the initial mesh of the anatomical surface; construct a model of the anatomical surface from the set of surface points and the set of interior points; andoutput a graphical representation of the model of the anatomical surface to the display.
15. The electroanatomical mapping system according to claim 14, wherein the plurality of ultrasound images comprises a plurality of two-dimensional ultrasound images.Attorney Docket No. 16043 WOO 1 / 82410- 131916. The electroanatomical mapping system according to claim 14, wherein the surface modeling module is configured to generate the set of surface points from the plurality of ultrasound images by executing a process comprising:receiving a user input identifying a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; anddiscretizing the identified blood pool / tissue boundary of the anatomical volume into the set of surface points.
17. The electroanatomical mapping system according to claim 14, wherein the surface modeling module is configured to generate the set of surface points from the plurality of ultrasound images by executing a process comprising:segmenting a blood pool / tissue boundary of the anatomical volume in the plurality of ultrasound images; anddiscretizing the segmented blood pool / tissue boundary of the anatomical volume into the set of surface points.
18. The electroanatomical mapping system according to claim 14, wherein the surface modeling module is configured to construct the initial mesh of an anatomical surface of the anatomical volume from the set of surface points by executing a process comprising:estimating normals for the set of surface points; andgenerating the initial mesh of the anatomical surface using a Poisson surface reconstruction algorithm.
19. The electroanatomical mapping system according to claim 18, wherein the surface modeling module is further configured to crop the initial mesh of the anatomical surface to a bounding box for the set of surface points to generate a cropped initial mesh.
20. The electroanatomical mapping system according to claim 18, wherein the surface modeling module is further configured to apply a hole-filling algorithm to the initial mesh of the anatomical surface to generate a filled initial mesh.Attorney Docket No. 16043 WOO 1 / 82410- 131921. The electroanatomical mapping system according to claim 18, wherein the surface modeling module is further configured to resample the set of surface points across the initial mesh of the anatomical surface.
22. The electroanatomical mapping system according to claim 21, wherein the surface modeling module is configured to resample the set of surface points uniformly across the initial mesh of the anatomical surface.
23. The electroanatomical mapping system according to claim 14, wherein the surface modeling module is configured to identify the set of interior points using the initial mesh of the anatomical surface by executing a process comprising:generating a plurality of points interior to a bounding box for the initial mesh; identifying a subset of the plurality of points that are interior to both the initial mesh and a convex hull of the set of surface points; anddefining the subset of the plurality of points as the set of interior points.
24. The electroanatomical mapping system according to claim 23, wherein the plurality of points interior to the bounding box for the initial mesh are randomly distributed within the bounding box for the initial mesh.
25. The electroanatomical mapping system according to claim 14, wherein the modeling module is configured to construct the model of the anatomical surface from the set of surface points and the set of interior points by applying an alpha shape algorithm to the set of surface points and the set of interior points.
26. The electroanatomical mapping system according to claim 14, wherein the modeling module is configured to construct the initial mesh of the anatomical surface from the set of surface points by applying a patch geometry algorithm to the set of surface points.
27. A computer-implemented method of modeling an anatomical surface bounding an anatomical volume, the method comprising:receiving a plurality of ultrasound images of the anatomical volume in an electroanatomical mapping system, the electroanatomical mapping system including a display;Attorney Docket No. 16043 WOO 1 / 82410- 1319via the electroanatomical mapping system:generating a set of surface points from the plurality of ultrasound images; constructing an initial mesh of the anatomical surface from the set of surface points;cropping the initial mesh of the anatomical surface to a bounding box for the set of surface points to create a cropped initial mesh;filling holes within the cropped initial mesh to create a filled initial mesh; identifying a set of interior points using the filled initial mesh; and constructing a model of the anatomical surface from the set of surface points and the set of interior points; andoutputting a graphical representation of the model of the anatomical surface to the display.
