Construction of Luminal Wall Topography in 4D Ultrasound Images Using Virtual Ellipsoids or Polyhedrons

Abstract

The method includes receiving a 3D image of a portion of an organ including a cavity, the image including voxels having respective values, where walls of the cavity are identified in the image by (i) positioning a virtual solid object within a respective sub-volume of the cavity, (ii) moving the virtual solid object within the cavity according to predefined rules of motion, (iii) adapting voxels traversed while the virtual solid object moves within the cavity to predefined values ​​indicative of an interior of the cavity, (iv) in response to detecting contact of the virtual solid object with a wall of the cavity, rolling the virtual solid object on the wall and adapting a surface of the virtual solid object to values ​​of the voxels over which the surface rolls, and (v) converting adapted voxel values ​​below a threshold voxel value to the predefined value indicative of an interior.

Description

[Technical field]

[0001] FIELD OF THE DISCLOSURE This disclosure relates generally to image processing, and specifically to image processing of ultrasound images. [Background technology]

[0002] Techniques for generating visual representations of portions of organs have been previously proposed in the patent literature. For example, U.S. Pat. No. 10,376,320 describes devices, systems, and methods aimed at generating three-dimensional surface representations of anatomical structures, such as heart chambers. More specifically, the three-dimensional surface representation of the anatomical structure is constrained to one or more anchor portions that correspond to received input regarding the location of anatomical features of the anatomical structure. To this end, the three-dimensional surface representation can be extracted from the three-dimensional data structure according to any one or more computational algorithms known in the art for volumetrically smoothing three-dimensional representations of objects, including the "ball pivot" algorithm, the "power crust" algorithm, and other similar algorithms. The resulting three-dimensional surface representation includes the salient features of the anatomical structure and can therefore be useful as a visualization tool during any of a variety of different medical procedures, including, for example, cardiac ablation.

[0003] As another example, US Patent Application Publication No. 2020 / 0211280 describes image processing performed by accepting an array of voxels containing data representing physical properties of a three-dimensional object, segmenting the array of voxels into a plurality of regional subarrays of voxels each meeting a predetermined criterion, converting the subarrays into respective triangular meshes, the meshes having triangles that enclose the subarray and occlude voxels outside the subarray, and rendering the triangular meshes on a display. In one embodiment, the processor generates the mesh using a Ball-Pivoting Algorithm (BPA). Typically, when BPA is used, the size of the ball is set to match the size of the voxel described above.

[0004] The present disclosure will become more fully understood from the following detailed description of the embodiments thereof, taken in conjunction with the drawings, in which: [Brief description of the drawings]

[0005] [Figure 1] 1 is a schematic, pictorial illustration of a catheter-based system configured to perform 4D ultrasound imaging of a heart chamber, according to one embodiment of the present disclosure. FIG. [Figure 2A] 1 is a schematic, pictorial illustration of cavity construction by the disclosed image processing techniques applied to 3D ultrasound images, according to some embodiments of the present disclosure. [Figure 2B] 1 is a schematic, pictorial illustration of cavity construction by the disclosed image processing techniques applied to 3D ultrasound images, according to some embodiments of the present disclosure. [Figure 3A] 1A-1C are schematic depictions of intermediate cavity construction by shell and by rebound path, respectively, according to some embodiments of the present disclosure; [Figure 3B] 1A-1C are schematic depictions of intermediate cavity construction by shell and by rebound path, respectively, according to some embodiments of the present disclosure; [Figure 4] 1 is a flow chart that generally illustrates a method for constructing a cavity by the disclosed image processing techniques applied to a 3D ultrasound image, according to one embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0006] overview Three-dimensional (3D) or four-dimensional (4D) ultrasound images of an organ may be acquired non-invasively or by a probe such as a catheter using a two-dimensional (2D) array of ultrasound transducers. In this context, the term "3D ultrasound image" refers to an ultrasound image that represents a particular volume in three dimensions. The term "4D ultrasound image" refers to a time series of 3D ultrasound images of a particular volume. A 4D image can be considered as a 3D movie with the fourth dimension being time. Another way to describe a 4D image (or rendering) is as a time-dependent 3D image (or rendering).

[0007] Physicians often wish to inspect the tissue walls of organ cavities (e.g., lumens) in 3D images in order to visualize and present the visualized luminal walls captured in the image. Optionally, a model (or surface) of the lumen may be constructed based on the visualization for display to the physician. If the organ lumen is empty or filled with liquid, it may be easy to distinguish voxels (image voxels) belonging to the tissue of the lumen wall from voxels belonging to the volume within the lumen. However, if the lumen is filled with blood (or other material), the blood (or other material) may also appear gray and may be difficult to visualize, e.g., the blood (or other material) may add noise to the image. This problem is intensified in cardiac images due to the movement of the blood and ventricles being imaged.

[0008] The embodiments of the disclosure described herein provide methods and systems applied to medical images, such as 3D ultrasound images, to construct organ cavities by finding cavity voxels (blood voxels in the case of a heart chamber) and then setting the cavity voxels with predefined cavity voxel values ​​(e.g., given Hounsfield unit (HU) values ​​that appear black in an ultrasound image) that are easily distinguishable from wall tissue voxel values. The result is a clear depiction of the organ cavity wall tissues displayed to the physician (e.g., in a 3D ultrasound rendering), including those within the heart chambers.

[0009] In some examples, an algorithm is provided that is applied by the processor to analyze the 3D image by first creating a virtual solid object, such as an ellipsoid (e.g., a virtual ball) or a virtual polyhedron (e.g., a virtual cube) within the cavity. The processor moves the virtual object within the cavity volume using the laws of motion described below. When engaging a surface, for example, the object rolls across and / or bounces off the surface to gradually build up the surface of the lumen wall.

[0010] Since the surfaces of organ cavities, such as the cardiac cavities of the heart, are typically curved, in some cases an ellipsoid is a good choice as a shape for tracing the surface. The dimensions of the ellipsoid, such as the diameter of the ball, can be defined based on the desired resolution and expected geometry of the organ. In other cases, for example, to best geometrically fit the virtual object shape to the 3D image voxels, the processor creates a virtual cube, which may be as small as a single voxel. The virtual cube, which moves within the cavity volume when engaging the surface, "rolls" around the 3D image facet by facet (e.g., voxel by voxel) to build the surface of the lumen wall with the necessary modifications in a manner similar to that done with a ball. In some instances, another virtual object shape, such as a polyhedron (e.g., an icosahedron) with triangular facets, may be found to be most suitable.

[0011] In one example, the processor begins the cavity construction process by positioning a virtual ball (or polyhedron) within a subvolume that is entirely within the volume known to be the inner lumen of the cavity (i.e., not touching any walls). The subvolume is known to be such by, for example, its voxel values, which correspond to the black color of the fully immersed subvolume.

[0012] As the ball (or polyhedron) rolls, the processor matches pixel values ​​(e.g., color) of the ball's (or polyhedron's) surface to the voxel values ​​of the organ interior wall that the ball touches. Initially, the ball's surface is completely black, since it is within a completely black volume. The rolling path can begin as a random, pseudo-random, or planned path. Continuing with the ball as an example, the rolling ball will match the ball's surface color to the colors of the voxels it touches, e.g., the surface of a virtual ball is "painted" with the image voxel values ​​that the virtual ball touches. The following steps are performed:

[0013] If the entire surface of the ball is black, then the ball is located in a portion of the 3D image that represents the lumen. The processor continues to move the ball and check the color.

[0014] If the ball surface (or a portion thereof) contains scattered grey or white speckles that do not form a continuous blob, the portion of the ball containing the scattered speckles is considered to still be located in the portion of the 3D image representing the lumen. The speckles are typically the result of blood flowing through the lumen. The processor changes all of the image voxels at that location to black accordingly. This removes the speckle noise caused by blood flow. The processor continues to move the ball and check the color.

[0015] If a portion of the ball's surface contains a white / gray patch, the processor considers the patch to be part of a tissue wall. The boundary is defined by setting a threshold such that a voxel value greater than the threshold means that the voxel is part of the cavity wall tissue. For example, the threshold voxel value may correspond to the color of a contiguous sphere of cavity wall tissue of a predefined size. The processor marks these image voxels as tissue walls (using white or some other color).

[0016] In one example, the ball rolls in a random direction, while in another example, the ball mat rolls in a pseudo-random manner to minimize the chance of the ball repeating its track.

[0017] Once the color on the ball has been checked, the processor can begin to roll the clean ball, etc., until all surface voxels have been accounted for. In this example, the processor has any moved cavity voxels that match a predefined cavity voxel value (e.g., black). This creates a black shell surrounding the reduced cavity. The processor then changes the color of each voxel in the reduced cavity to a value indicative of interior, such as black (i.e., converts the values ​​of all cavity voxels inside the shell to cavity voxel values), thereby achieving a perfectly well-defined black cavity bounded by a clearly delineated cavity wall.

[0018] In another example, when the color on the ball is checked, the processor changes the direction of the ball so that it bounces off the detected tissue wall. Also in this example, the processor has any displaced cavity voxels that match a predefined cavity voxel value (e.g., black). However, the cavity color is formed during the step where the processor starts to roll the clean ball at the wall tissue location. As the ball bounces, both the tissue wall surface and cavity color are defined. The bounce path can be initiated as random or pseudorandom, for example, into a posterior hemisphere distribution of directions. Alternatively, the bounce path may be along a planned path or a path governed by physical rules (e.g., specular reflection). Note that any physical rules used by the algorithm need not reflect the exact physical low of the motion. For example, the processor can preserve the specular direction of the bounce even though the ball has already rolled away from where it encountered the wall tissue.

[0019] In yet another example, a ball is positioned in a volume that is entirely black, or in a volume known to be a lumen. The ball's surface simply bounces between the lumen walls, adapting to the color it comes into contact with. Initially, the ball's surface is entirely black. The bounce path can begin as a random, pseudorandom, planned path, or a path governed by physical rules (e.g., specular reflection). As the ball moves and hits a lumen wall location, the surface color is updated to match the color the ball is hitting, e.g., the virtual ball's surface is "painted" with the image voxel value where the ball is hitting. The color is compared to a threshold as described above. At the same time, the processor has any moved lumen voxels that match a predefined lumen voxel value (e.g., black).

[0020] The process continues until all candidate surface voxels have been considered (e.g., touched or passed through). Alternatively, or in parallel, the algorithm stops if it cannot find a new sub-volume location for the initial ball placement.

[0021] The spatial resolution of the surface depends on the diameter of the ball. A small virtual ball provides greater resolution but requires more time to build the tissue wall surface. A large virtual ball provides less resolution but builds the tissue wall surface more quickly. Such a ball may be suitable for use with a large cavity or may be suitable for a first build round before a build round with a smaller diameter ball. In one example, the diameter of the ball may be detected and the process may present to the user to decide whether to modify the diameter.

[0022] In some examples, the processor divides the image volume into multiple blocks and runs a virtual ball through each block. In this way, all of the balls can be in motion simultaneously (using parallel processing). The rendered surfaces of each block are connected when the process is finished, so that the surface portions constructed in each of the blocks are combined to construct the entire surface. In one example, the number and / or size of the balls are user selected.

[0023] System Description 1 is a schematic, pictorial illustration of a catheter-based ultrasound imaging system 20 using a catheter 21 with a distal tip assembly 40 comprising a 2D ultrasound array 50 and a position sensor 52, according to one embodiment of the present disclosure. The integrated location sensor 52 is pre-aligned with the 2D array 50 of the catheter 21.

[0024] In particular, the sensor 52 is configured to output signals indicative of the position, direction and orientation of the 2D ultrasound transducer array 52 within the organ. The system's processor is configured to use the signals output by the sensor to register multiple ultrasound image sections acquired by the 2D ultrasound transducer array 50 with respect to one another.

[0025] As can be seen, the distal tip assembly 40 is attached to the distal end of the catheter shaft 22. The catheter 21 is inserted through a sheath 23 and into a heart 26 of a patient 28 lying on a surgical table 29. The proximal end of the catheter 21 is connected to a control console 24. In the embodiments described herein, the catheter 21 is used for ultrasound-based diagnostic purposes, however, the catheter may further be used to perform therapy, such as electrical sensing and / or ablation of tissue within the heart 26, for example, using the tip electrode 56.

[0026] A physician 30 navigates a distal end assembly 40 of catheter 21 to a target location within heart 26 by manipulating shaft 22 using a manipulator 32 near the proximal end of the catheter.

[0027] In one embodiment, a 2D ultrasound array 50, shown in detail in inset 25, is configured to image the left atrium of heart 26. Recorded images are stored by processor 39 in memory 37.

[0028] As can be seen in inset 45, the ultrasound array 50 includes a 2D array 50 of multiple ultrasound transducers 53. Inset 45 shows the ultrasound array 50 navigated relative to the ostial wall 54 of a pulmonary vein in the left atrium. In this example, the 2D array 50 is an array of 32x64 US transducers. The 2D array is capable of imaging a section of the inner wall of the ostium. Due to the integrated location sensor, the spatial coordinates of all pixels within the imaged section are known.

[0029] The control console 24 includes a processor 39, typically a general purpose computer, along with suitable front end and interface circuitry 38 for driving the ultrasound transducers 53 (e.g., in a phased array manner including steering the ultrasound beam) and for receiving echo signals from the transducers 53 for use by the processor 39. The interface circuitry 38 receives signals from the catheters 21 and is optionally further used to deliver therapy via the catheters 21 within the heart 26 and to control other components of the system 20. The console 24 also includes a driver circuit 34 configured to drive the magnetic field generators 36.

[0030] During navigation of the distal end 22 in the heart 26, the console 24 receives position and orientation signals from the location sensor 52 in response to a magnetic field from an external magnetic field generator 36. The magnetic field generator 36 is disposed at a known location outside the patient 28, for example, beneath the table 29 on which the patient lies. These position and orientation signals indicate the position and orientation of the 2D ultrasound array 50 in the coordinate system of the position tracking system.

[0031] Position and orientation sensing methods using external magnetic fields have been implemented in a variety of medical applications, for example in the CARTO™ system manufactured by Biosense Webster, and are described in detail in U.S. Pat. Nos. 6,618,612 and 6,332,089, WO 96 / 05768, and U.S. Patent Application Publication Nos. 2002 / 0065455, 2003 / 0120150, and 2004 / 0068178, the disclosures of which are all incorporated herein by reference.

[0032] In some embodiments, the processor 39 may be configured to operate the array 52 in a "sweeping mode" to image a large portion of the cardiac camber. In one example, the ultrasound imaged heart chamber (e.g., ventricle) is presented by the processor 39 to the physician 30 on the monitor 27, e.g., as a volume rendering 55, where the cavity 57 appearing in the rendering 55 has been reconstructed by the processor 39 using the disclosed technique of constructing a topography of the lumen wall as a 4D ultrasound image with a virtual ellipsoid or polyhedron, as further described in FIG.

[0033] The processor 39 typically comprises a general-purpose computer that is programmed with software to perform the functions described herein. The software may be downloaded into the computer in electronic form, for example over a network, or alternatively or additionally may be provided and / or stored on a non-transitory tangible medium, such as magnetic, optical, or electronic memory.

[0034] The exemplary configuration shown in FIG. 1 is selected solely for purposes of conceptual clarity. The techniques of the present disclosure can be similarly applied using other system components and configurations. For example, system 20 may include additional components and perform non-cardiac catheterization procedures. In particular, processor 39 executes the dedicated algorithms disclosed herein, including FIG. 4, that enable processor 39 to perform the disclosed steps, as described further below.

[0035] Construction of Luminal Wall Topography in 4D Ultrasound Images Using One or More Virtual Balls As described above, in one example, a processor such as processor 39 receives a three-dimensional (3D) image of at least a portion of an organ including a cavity, the image including voxels having respective values. The processor identifies the walls of the cavity in the image by applying the following steps to the image: 1. Positioning one or more virtual solid objects of a predefined size and shape within one or more respective sub-volumes of the cavity, each enclosed within the cavity. 2. Moving one or more virtual solid objects within the cavity according to predefined rules of motion; Adapting voxels traversed by the one or more virtual solid objects while the one or more virtual solid objects move within the cavity to a predefined value indicative of the interior of the cavity. 3. In response to detecting contact of the virtual solid object with a wall of the cavity, rolling the virtual solid object on the wall and adapting a surface of the virtual solid object to the values ​​of the voxels over which the surface rolls. 4. Converting the fitted voxel values ​​below a threshold voxel value to a predefined value that indicates interior.

[0036] 2A and 2B are schematic, pictorial illustrations of cavity construction by the disclosed image processing techniques applied to a 3D ultrasound image 202, according to some embodiments of the present disclosure.

[0037] FIG. 2A shows a ball 205 of a given diameter, and as the ball 205 rolls over the wall 210 of the cavity 204, its surface 206 adapts the voxel values ​​(e.g., color) of the cavity wall it contacts. It should be noted that the ball 205 may be fully virtual, i.e., constructed within the virtual environment of the processing system, for the purpose of generating a cleaner view for presentation on a display for the physician and / or operator. In other words, the virtual ball may not be visible to the user on any display (although it may be visible if desired). Thus, the depiction of the exemplary ball 205 in this disclosure should be considered primarily for the purpose of illustrating the technique of generating a clean view of the target anatomical structure.

[0038] The rolling path can begin as a random, pseudo-random, or planned path. The surface of the virtual ball is "painted" with contiguous image voxel values ​​according to the following steps: If the entire surface of the ball is black (not shown in this case), the ball is located in a portion of the 3D image that represents the lumen. The processor continues to move the ball and check the color.

[0039] If the entire ball surface 206 (or a portion of it) contains scattered grey or white speckles that do not form a continuous blob, the portion of the ball containing the scattered speckles is considered to still be located in a portion of the 3D image representing the lumen. The processor changes all of the image voxels at that location to black accordingly. This removes the "snow" noise caused by blood flow. The processor continues to roll the ball and checks the color.

[0040] If a portion of the ball's surface contains a white / gray patch, the patch is part of a tissue wall. The boundary is defined by setting a threshold, so that a voxel value greater than the threshold means that the voxel is part of the cavity wall tissue. The processor marks these tissue wall image voxels with white or some other light gray shade. The processor then changes the direction of the ball so that it bounces off the detected tissue wall (not shown).

[0041] In Fig. 2B, which shows an alternative or complementary process to that described above, the processor divides the volume of the image 202 into two blocks 212 and 214 and runs a virtual ball within each block. A large diameter ball 216 is used in the larger image volume on the right, while a smaller diameter ball 218 is used in the smaller cavity volume. In this way, the ball diameter can be optimized for the size of the portion of the volume. A smaller diameter ball is suitable, for example, for smaller cavities or to achieve a higher surface resolution.

[0042] Furthermore, since two balls are manipulated simultaneously (using parallel processing), this facilitates cavity construction.

[0043] At the end of this procedure, the surfaces of each of the blocks 212 and 214 are connected such that the surface portions constructed on each of the blocks can be joined to construct the entire surface.

[0044] 3A and 3B are schematic depictions of an intermediate cavity structure 302 with a shell 306 and a structure 312 with a bounce path 322 (eg, a track 322), respectively, according to some embodiments of the present disclosure.

[0045] 3A, a ball such as ball 218 is rolled by processor 39 over the cavity wall (e.g., wall 210) and as the ball rolls (e.g., passes) through the cavity voxels, they match predefined cavity voxel values ​​(e.g., black). This creates a black shell 306 that surrounds the boundary 380 of the reduced cavity 310. The processor then changes the color of each voxel of the reduced cavity 310 to black (i.e., converts the values ​​of all cavity voxels inside shell 306 to cavity voxel values), thereby achieving a perfectly well-defined black cavity that is bounded by a clearly delineated cavity wall 304.

[0046] In Fig. 3B, the ball bounced off the detected tissue wall traveled in a path 322 inside the cavity, passing through cavity voxels in the path 322 and adapting predefined cavity voxel values ​​(e.g., black). In this example, the ball bounces between (and adapts the color of) the locations where it hits on the cavity wall 314. The bouncing path may be along a planned path or a path governed by physical rules (e.g., specular reflection).

[0047] The construction steps described in Figures 2A-2B and 3A-3B are provided by way of example and have been simplified for clarity of presentation. As noted above, additional or alternative steps may be performed.

[0048] Method for constructing cavities in 4D ultrasound images by manipulating virtual objects 4 is a flow chart that illustrates, in accordance with one embodiment of the present disclosure, a method for constructing a cavity by the disclosed image processing techniques applied to a 3D ultrasound image. The algorithm according to the presented embodiment performs a process that begins with the processor 39 receiving, in step 402, a 3D ultrasound image of at least a portion of an organ that includes a cavity, the image including voxels with respective values.

[0049] The processor executes the disclosed algorithm, in a virtual object positioning step 404, by positioning one or more virtual solid objects of a predefined size and shape within one or more respective sub-volumes of the cavity, each entirely within a volume known to be that of the cavity (e.g., as shown in FIG. 2A or FIG. 2B).

[0050] Next, the processor 39, in an object movement step 406, moves each of the one or more virtual solid objects within the cavity according to a predefined rule of movement, such as having a linear trajectory and bouncing off the cavity wall in a specular direction (after completing its roll against the wall) in a manner similar to the motion of a billiard ball. Alternatively, the processor may apply a random rule (e.g., a distribution of directions with probabilities, where the direction is chosen randomly by an algorithm) to the direction of bouncing off the cavity wall.

[0051] The processor periodically checks, in a checking step 408, whether each object has encountered a wall.

[0052] If the answer is no, then one or more virtual solid objects are free to move within the cavity and the processor matches any passing cavity voxels to predefined cavity voxel values ​​in a conversion to cavity voxels step 410, and the processor loops back to step 406.

[0053] If the answer is yes, i.e., if one or more virtual solid objects each come into contact with the cavity wall, then in an object rolling step 412, processor 39 rolls the virtual solid objects against the cavity wall tissue so that the surfaces of the virtual solid objects match the cavity voxel tissue values ​​against which they are rolled.

[0054] In a voxel value conversion step 414, the processor converts the matched voxel values ​​that are below a threshold voxel value to a predetermined cavity voxel value so as to reconstruct the cavity voxels using only the cavity voxel values.

[0055] In a termination criteria check step 416, the processor 39 checks whether all initial positions (i.e., completely enclosed cavity positions) have been used. If the answer is no, the processor returns to step 404. An alternative termination criterion checked in step 36 may be whether all candidate surface voxels have been considered. If the answer is no, the processor returns to step 404.

[0056] If the answer is yes, then the processor, in a Present Processed 3D Ultrasound Image step 418, presents on the display a denoised 3D image that clearly shows the cavity and its walls.

[0057] The flow chart of Figure 4 is an example. Images other than ultrasound images (e.g., noisy CT images) can be similarly processed to make the cavity more visible. The construction method described in Figure 4 is provided by way of example and is simplified for clarity of presentation. As noted above, additional steps, partial steps (e.g., only bouncing or only rolling), or alternative steps may be performed. EXAMPLES

[0058] The method includes receiving a three-dimensional (3D) image (202) of at least a portion of an organ including a cavity (204), the image including voxels having respective values ​​of a wall (210) of the cavity in the image (202), (i) positioning one or more virtual solid objects (205, 216, 218) of a predefined size and shape within one or more respective sub-volumes of the cavity (204) each enclosed within the cavity, (ii) moving the one or more virtual solid objects (205, 216, 218) within the cavity according to predefined rules of motion, and (iii) visualizing the one or more virtual solid objects (205, 216, 218) as they move within the cavity (204). (iv) in response to detecting the virtual solid object contacting a wall of the cavity, rolling the virtual solid object over the wall (210) and adapting the surface (206) of the virtual solid object to values ​​of the voxels over which the surface (206) rolls; and (v) converting adapted voxel values ​​that are below a threshold voxel value to a predefined value that is indicative of the interior. EXAMPLES

[0059] 2. The method of claim 1, wherein the predefined value indicating the interior corresponds to the color black in the 3D image (202). EXAMPLES

[0060] 3. The method according to any of the preceding claims, wherein the threshold voxel value corresponds to the colour of a contiguous sphere of a predefined size of tissue of the wall (210) of the cavity (204). EXAMPLES

[0061] The method according to any of the preceding claims, wherein the predefined motion rules specify that one or more virtual solid objects (205, 216, 218) should move in a linear trajectory (322) within the cavity. EXAMPLES

[0062] The method according to any of the preceding claims, wherein the predefined motion rules specify bouncing one or more virtual solid objects (205, 216, 218) off the walls (210) of the cavity (204) in a specular direction. EXAMPLES

[0063] The method according to any of the preceding claims, wherein the predefined movement rules only allow the one or more virtual solid objects (205, 216, 218) to roll on the walls (210) of the cavity (204). EXAMPLES

[0064] 7. The method according to any of claims 1 to 6, comprising reconstructing voxels belonging to the interior by defining a shell (306) along the walls (314) and converting values ​​of voxels inside the shell to values ​​indicative of the interior. EXAMPLES

[0065] The method according to any one of Examples 1 to 7, wherein the predefined movement rule is a random movement. EXAMPLES

[0066] The method according to any one of the preceding embodiments, wherein the virtual solid object (304) is an ellipsoid. EXAMPLES

[0067] The method according to any one of the preceding embodiments, wherein the virtual solid object (304) is a polyhedron. EXAMPLES

[0068] The method according to any one of the preceding claims, wherein the 3D image (202) is a 3D ultrasound image. EXAMPLES

[0069] The method according to any one of the preceding claims, wherein the cavity (204) is a cardiac cavity. EXAMPLES

[0070] The method according to any of the preceding embodiments, wherein the voxel values ​​are given in Hounsfield Units (HU). EXAMPLES

[0071] The system includes a memory (37) and a processor (39). The memory (37) is configured to store a three-dimensional (3D) image (202). The processor (39) is configured to receive the 3D image (202) of at least a portion of an organ including a cavity (204), the image including voxels having respective values, and includes the steps of (i) positioning one or more virtual solid objects (205, 216, 218) of a predefined size and shape within one or more respective sub-volumes of the cavity (204) each enclosed within the cavity, (ii) moving the one or more virtual solid objects (205, 216, 218) within the cavity according to predefined rules of motion, and (iii) displaying the one or more virtual solid objects (205, 216, 218) in accordance with the predefined rules of motion. The wall (210) of the cavity is identified in the image (202) by applying to the image the steps of: (i) adapting voxels (322) traversed by the virtual solid object while the virtual solid object (216, 218) moves within the cavity (204) to predefined values ​​indicative of the interior of the cavity; (iv) in response to detecting that the virtual solid object contacts the wall of the cavity, rolling the virtual solid object over the wall (210) and adapting the surface (206) of the virtual solid object to the values ​​of the voxels over which the surface (206) rolls; and (v) converting adapted voxel values ​​that are below a threshold voxel value to a predefined value indicative of the interior.

[0072] Although the examples described herein primarily address cardiac applications, the methods and systems described herein may also be used in other body organs, for example, the disclosed techniques may be used for ultrasound imaging of the stomach and bladder.

[0073] Therefore, it will be understood that the above-described embodiments are given by way of example, and that the present disclosure is not limited to what has been particularly shown and described above. Rather, the scope of the present disclosure includes both combinations and subcombinations of the various features described above, as well as variations and modifications thereof not disclosed in the prior art that would occur to one skilled in the art upon reading the foregoing description. Documents incorporated by reference into this patent application shall be considered as an integral part of this application, provided that, to the extent that any term is defined in these incorporated documents in a manner that is inconsistent with the definition expressly or implicitly made herein, only the definition in this specification shall be considered.

[0074] [Embodiment] (1) A method comprising: receiving a three-dimensional (3D) image of at least a portion of an organ including a cavity, the image including voxels having respective values; identifying a wall of the cavity in the image, positioning one or more virtual solid objects of a predefined size and shape within one or more respective sub-volumes of the cavity, each sub-volume being enclosed within the cavity; moving the one or more virtual solid objects within the cavity according to predefined rules of motion; adapting the voxels traversed by the one or more virtual solid objects while the one or more virtual solid objects move within the cavity to a predefined value indicative of the interior of the cavity; in response to detecting contact of a virtual solid object with a wall of the cavity, rolling the virtual solid object over the wall and conforming a surface of the virtual solid object to the values ​​of the voxels over which the surface is rolled; and converting the matched voxel values ​​that are below a threshold voxel value to the predefined value that indicates the interior by applying the predefined value to the image. (2) The method of claim 1, wherein the predefined value indicating the interior corresponds to black in the 3D image. (3) The method of embodiment 1, wherein the threshold voxel value corresponds to the color of a contiguous sphere of a predefined size of tissue of the wall of the cavity. (4) The method of embodiment 1, wherein the predefined motion rules specify that the one or more virtual solid objects are to move in a linear trajectory within the cavity. (5) The method of embodiment 1, wherein the predefined motion rules specify bouncing the one or more virtual solid objects off the walls of the cavity in a specular direction.

[0075] (6) The method of embodiment 1, wherein the predefined movement rules only allow the one or more virtual solid objects to roll on the walls of the cavity. (7) A method as described in embodiment 1, comprising defining a shell along the wall and reconstructing the voxels belonging to the interior by converting the values ​​of the voxels inside the shell to values ​​indicative of the interior. (8) The method of embodiment 1, wherein the predefined movement rule is a random movement. (9) The method of embodiment 1, wherein the virtual solid object is an ellipsoid. (10) The method of embodiment 1, wherein the virtual solid object is a polyhedron.

[0076] (11) The method of embodiment 1, wherein the 3D image is a 3D ultrasound image. (12) The method of embodiment 1, wherein the cavity is a cardiac cavity. (13) The method of embodiment 1, wherein the value of the voxel is given in Hounsfield Units (HU). (14) A system comprising: a memory configured to store a three-dimensional (3D) image; a processor, the processor comprising: receiving a 3D image of at least a portion of an organ including a cavity, the image including voxels having respective values; identifying a wall of the cavity in the image, positioning one or more virtual solid objects of a predefined size and shape within one or more respective sub-volumes of the cavity, each sub-volume being enclosed within the cavity; moving the one or more virtual solid objects within the cavity according to predefined rules of motion; adapting the voxels traversed by the one or more virtual solid objects while the one or more virtual solid objects move within the cavity to a predefined value indicative of the interior of the cavity; in response to detecting contact of a virtual solid object with a wall of the cavity, rolling the virtual solid object over the wall and adapting a surface of the virtual solid object to the values ​​of the voxels over which the surface is rolled; and converting the adapted voxel values ​​that are below a threshold voxel value to the predefined value that indicates the interior by applying a threshold voxel value to the image. (15) The system of embodiment 14, wherein the predefined value indicating the interior corresponds to black in the 3D image.

[0077] (16) The system of embodiment 14, wherein the threshold voxel value corresponds to a color of a contiguous sphere of a predefined size of tissue of the wall of the cavity. (17) The system of embodiment 14, wherein the predefined motion rules specify that the one or more virtual solid objects are to move in a linear trajectory within the cavity. (18) The system of embodiment 14, wherein the predefined motion rules specify bouncing the one or more virtual solid objects off the walls of the cavity in a specular direction. (19) The system of embodiment 14, wherein the predefined movement rules only allow the one or more virtual solid objects to roll on the walls of the cavity. (20) The system of claim 14, wherein the processor is further configured to define a shell along the wall and reconstruct the voxels belonging to the interior by converting the values ​​of the voxels inside the shell to values ​​indicative of the interior.

[0078] (21) The system described in embodiment 14, wherein the predefined movement rule is a random movement. (22) The system of embodiment 14, wherein the virtual solid object is an ellipsoid. (23) The system of embodiment 14, wherein the virtual solid object is a polyhedron. (24) The system of embodiment 14, wherein the 3D image is a 3D ultrasound image. (25) The system of embodiment 14, wherein the cavity is a heart cavity.

[0079] (26) The system of embodiment 14, wherein the value of the voxel is given in Hounsfield Units (HU).

Claims

1. 1. A system comprising: a memory configured to store a three-dimensional (3D) image; a processor, the processor comprising: receiving a 3D image of at least a portion of an organ including a cavity, the image including voxels having respective values; identifying a wall of the cavity in the image, positioning one or more virtual solid objects of a predefined size and shape within one or more respective sub-volumes of the cavity, each sub-volume being enclosed within the cavity; moving the one or more virtual solid objects within the cavity according to predefined rules of motion; adapting the voxels traversed by the one or more virtual solid objects while the one or more virtual solid objects move within the cavity to a predefined value indicative of the interior of the cavity; in response to detecting contact of a virtual solid object with a wall of the cavity, rolling the virtual solid object over the wall and conforming a surface of the virtual solid object to the values ​​of the voxels over which the surface is rolled; and converting the adapted voxel values ​​that are below a threshold voxel value to the predefined value that indicates the interior by applying a threshold voxel value to the image.

2. The system of claim 1 , wherein the predefined value indicating interior corresponds to black in the 3D image.

3. The system of claim 1 , wherein the threshold voxel value corresponds to a color of a contiguous sphere of a predefined size of tissue of the wall of the cavity.

4. The system of claim 1 , wherein the predefined motion rules specify that the one or more virtual solid objects are to move in a linear trajectory within the cavity.

5. The system of claim 1 , wherein the predefined motion rules specify bouncing the one or more virtual solid objects off the walls of the cavity in a specular direction.

6. The system of claim 1 , wherein the predefined movement rules only allow the one or more virtual solid objects to roll on the walls of the cavity.

7. 2. The system of claim 1, wherein the processor is further configured to reconstruct the voxels that belong to the interior by defining a shell along the walls and converting the values ​​of the voxels inside the shell to the values ​​that are indicative of the interior.

8. The system of claim 1 , wherein the predefined movement rule is a random movement.

9. The system of claim 1 , wherein the virtual solid object is an ellipsoid.

10. The system of claim 1 , wherein the virtual solid object is a polyhedron.

11. The system of claim 1 , wherein the 3D image is a 3D ultrasound image.

12. The system of claim 1 , wherein the chamber is a heart chamber.

13. The system of claim 1 , wherein the values ​​of the voxels are given in Hounsfield Units (HU).

14. 1. A method comprising: receiving a three-dimensional (3D) image of at least a portion of an organ including a cavity, the image including voxels having respective values; identifying a wall of the cavity in the image, positioning one or more virtual solid objects of a predefined size and shape within one or more respective sub-volumes of the cavity, each sub-volume being enclosed within the cavity; moving the one or more virtual solid objects within the cavity according to predefined rules of motion; adapting the voxels traversed by the one or more virtual solid objects while the one or more virtual solid objects move within the cavity to a predefined value indicative of the interior of the cavity; in response to detecting contact of a virtual solid object with a wall of the cavity, rolling the virtual solid object over the wall and conforming a surface of the virtual solid object to the values ​​of the voxels over which the surface is rolled; and converting the matched voxel values ​​that are below a threshold voxel value to the predefined value that indicates the interior by applying a predefined value to the image.

15. The method of claim 14 , wherein the predefined value indicating interior corresponds to the color black in the 3D image.

16. The method of claim 14 , wherein the threshold voxel value corresponds to a color of a contiguous sphere of a predefined size of tissue of the wall of the cavity.

17. The method of claim 14 , wherein the predefined motion rules specify that the one or more virtual solid objects are to move in a linear trajectory within the cavity.

18. The method of claim 14 , wherein the predefined motion rules specify bouncing the one or more virtual solid objects off the walls of the cavity in a specular direction.

19. The method of claim 14 , wherein the predefined movement rules only allow the one or more virtual solid objects to roll on the walls of the cavity.

20. 15. The method of claim 14, comprising: defining a shell along the walls and reconstructing the voxels that belong to the interior by transforming the values ​​of the voxels inside the shell to the values ​​that are indicative of the interior.

21. The method of claim 14 , wherein the predefined movement rule is a random movement.

22. The method of claim 14 , wherein the virtual solid object is an ellipsoid.

23. The method of claim 14 , wherein the virtual solid object is a polyhedron.

24. The method of claim 14 , wherein the 3D image is a 3D ultrasound image.

25. The method of claim 14 , wherein the chamber is a heart chamber.

26. The method of claim 14 , wherein the values ​​of the voxels are given in Hounsfield Units (HU).

Images