Computer-implemented method and system for navigating and displaying 3D image data

Abstract

A computer-implemented method and system for navigating and displaying 3D image data are described. In this method, a 3D image dataset to be displayed is retrieved, and highlighted locations are identified within the 3D image dataset. A scalar opacity map is calculated for the 3D image dataset, having values ​​for each of a plurality of locations in the 3D image dataset, each value depending on its position relative to the highlighted location and on the value of the 3D image at that location relative to the value of the 3D image at the highlighted location. The opacity is applied to the 3D image dataset to generate a modified 3D image view.

Description

Case Analysis

[0001] This application is a divisional application of the invention patent application filed on September 25, 2020, with application number 202080067305.8 and title "Computer-based method and system for navigating and displaying 3D image data". Technical Field

[0002] This invention relates to a system and computer-implemented method for navigating and displaying three-dimensional images, and is particularly applicable to three-dimensional imaging of human anatomy for medical diagnosis and treatment planning purposes. Background Technology

[0003] Traditional imaging scanners are used for a variety of purposes, including imaging the human and animal bodies for diagnosis and guidance during medical interventions such as surgery. Other uses of imaging scanners include structural analysis of buildings, pipes, and other structures.

[0004] Traditional medical ultrasound scanners create two-dimensional B-mode tissue images, where pixel brightness is based on the intensity of the echo return. Other types of imaging scanners can capture blood flow, tissue movement over time, blood location, the presence of specific molecules, tissue stiffness, or the anatomical structure of a three-dimensional (3D) region.

[0005] Traditionally, imaging scanners produce two-dimensional images. 2D images, such as 2D ultrasound images, cannot represent the typical 3D structure of human or animal organs because they only capture a 2D slice of a cross-section. However, if a probe, such as an ultrasound probe, scans the area of ​​interest mechanically or electronically, a three-dimensional image volume is generated. Alternatively, some ultrasound probes, such as “matrix” probes, have multiple piezoelectric crystals and can construct “real-time” 3D ultrasound images. This can then be displayed using, for example, 3D holography, and anatomical structures become easier to visualize for both trained and untrained observers because it better represents the true underlying structures / anatomical structures. Other techniques also allow for the capture or construction of 3D images.

[0006] In recent years, 3D imaging (in the form of ultrasound, CT, and MR) has been used by clinicians and has proven to be very valuable because it can convey imaging information in an intuitive format. For example, in the field of cardiology, such data is used to plan and guide surgical procedures and catheter interventions.

[0007] The current limitation of 3D imaging is that, although the data is inherently 3D, traditional 2D displays can only render a planar representation of the image on the screen (projection, slicing, projection, etc.).

[0008] As mentioned above, 3D imaging devices are available. However, most computing systems, including those in imaging systems, have two-dimensional displays and user interface designs for two-dimensional navigation. Until recently, technologies using computational reality techniques such as holograms, virtual reality, mixed reality, or augmented reality to display 3D images have become available. However, the development of such technologies has not primarily addressed the specific requirements of clinical settings. 3D systems tend to be expensive, and their interfaces are unfamiliar to users accustomed to working in 2D.

[0009] Another problem with 3D data rendering is the significant increase in the amount of information depicted to the user.

[0010] While this can be considered a positive development, it also makes navigating and changing views in three-dimensional space, as well as absorbing information about features of interest, more difficult.

[0011] Typically, these problems mean that users revert to using 2D displays and user interfaces to process 2D image data in two-dimensional slices. While this may be the user's preferred approach, it results in the loss of potentially relevant or interesting information from the 3D view (e.g., information from different orientations). Consequently, many advantages of 3D systems are lost, ultimately rendering 3D systems essentially expensive 2D systems. Summary of the Invention

[0012] According to one aspect of the present invention, a method and apparatus for navigating and displaying 3D image data are provided. The method includes:

[0013] Retrieve the dataset of 3D images to be displayed;

[0014] Receive the identifiers of the highlighted locations within the 3D image dataset;

[0015] Calculate a scalar opacity map of the 3D image dataset, the opacity map having a value at each of a plurality of locations in the 3D image dataset, the corresponding value depending on the corresponding location relative to the highlighted location, and depending on the value of the 3D image at the corresponding location relative to the value of the 3D image at the highlighted location; and

[0016] Opacity is applied to the 3D image dataset to generate a modified 3D image view.

[0017] In embodiments of the invention, a 3D image dataset refers to a 3D array of scalar or vector values, and the relevant range, orientation, and resolution that may allow for establishing correspondences between 3D images and objects in the real or imagined world. The methods and apparatus described herein, as well as the claims outlined, apply to this definition of 3D image data, but also to other definitions, including but not limited to non-Cartesian space sampling of 3D scalar or vector fields, such as 3D spherical sampling (e.g., for certain 3D ultrasound systems), 3D unstructured datasets (e.g., generated from computational fluid dynamics simulations), and point clouds (e.g., from particle image velocimetry). In all cases, the value of a point within a 3D image can be a color value or some other scalar or vector image-related value. This value may or may not have been captured by an image sensor and may be ultrasound, MRI, Doppler, or other data, but it may also represent detected blood velocities or other modalities or measurable / computable values ​​mapped to a 3D image.

[0018] In the case of multi-channel 3D image data, a variety of different methods can be used (which can be selected via a user interface or pre-selected based on the expected data). For example:

[0019] The system can apply the same opacity mask to multiple (not necessarily all) channels. It can calculate the opacity mask for a single channel, for multiple channels, and merge or calculate the opacity mask for flattened versions of the channels.

[0020] The system can selectively (by the system or the user) calculate and apply an opacity mask to only one channel;

[0021] For example, with color flow Doppler, the anatomical channel (B mode) can be faded out, but the blood flow (CFD channel) will not.

[0022] The system can apply opacity masks to multiple channels in a weighted manner (determined by the system or specified by the user via the user interface).

[0023] For example, for color flow Doppler, the anatomical channel (B mode) can fade out more than the blood flow fade-out (CFD channel), but less than the fade-out at the same distance from the highlight.

[0024] Alternatively, the fade-out distances of one channel and another can be different; for example, anatomical structures fade out very close to the highlight, while colors flow away from the highlight.

[0025] The computational steps may include using a masking kernel. Alternatively, a predefined shape or another 3D image can be used as the masking function.

[0026] The modified 3D image view can be rendered as a 2D or 3D image (or other rendering) to be displayed to the user.

[0027] Preferably, a scalar opacity map is calculated for a region of the 3D image dataset, which includes a portion of the 3D image dataset in the field of view between the highlighted location and the edge of the 3D image dataset.

[0028] An example of this type of 3D imaging is a 3D ultrasound image of the heart. These images include many structures around the heart that are opaque to ultrasound waves; thus, these structures obscure the view of the heart's internal structures. An example of this is... Figure 2a As shown in the diagram. Embodiments of the invention include methods and systems that enable a user to strip away blurred structures and focus on a structure of interest. Preferred embodiments have an intuitive user interface that defines the structure of interest using the center of a view from a designated highlight location within a 3D imaging volume. In this way, the user simply identifies the location and direction of the desired highlight (similar to shining a flashlight on an unlit scene), and the system is able to focus on the structure within that view. Preferably, a masking kernel is used (or the user can be given the ability to select one from multiple masking kernels), and the user interface includes user interface features through which the user can adjust parameters of the kernel. These parameters are preferably adjustable during use so that the user can change the degree to which surrounding structures are visible.

[0029] One kernel that can be used is the Gaussian kernel, which will be discussed below. Alternatively, other kernels will be understood, such as those based on a uniform / rectangular distribution, radial basis functions, spherical step functions, or exponential distribution (in the case of an exponential distribution, the user will select a point / region to obscure rather than highlight).

[0030] A preferred embodiment applies a position-dependent opacity kernel, causing the opacity of image features in the rendered 2D (or 3D) view of the 3D image dataset to change according to the position of the highlighted points. Preferably, the user interface allows the user to move the highlighted points and optionally use other parameters to control the opacity, as described in more detail below. Advantageously, an intuitive user interface is provided to the user for navigating the 3D image using a 2D display. Preferably, the user interface receives input from a keyboard and / or mouse and / or other controllers that interact with the user interface through the 2D display. In this way, the user can change the viewpoint / view around the 3D image and view the highlighted structures / areas from different perspectives. As a 3D image defined by voxels, the volume can be navigated and viewed using existing 3D rendering systems (or other renderings of 2D slices or 3D images).

[0031] Although the focus of the following discussion is on 3D images, it should be understood that embodiments of the invention are applicable to higher-dimensional datasets, such as 4D (3D images + time). In this case, the user interface may include the ability for the user to set points (or ranges) to be displayed in real time, or it may automatically cycle through recorded images for use in the view.

[0032] Similarly, dimensions do not need to correspond to (or fully correspond to) data from the visible spectrum, and can include representations of ultrasound, MRI (magnetic resonance imaging), or other data used to form a multispectral or spurious spectrum image to be viewed.

[0033] For example:

[0034] 3D Color Doppler Data

[0035] This modality consists of 3-channel 3D imaging data over time. Each time frame is a data volume, and for each voxel in the imaging data, there are two values ​​(channels): a background value corresponding to the B-mode (luminance) anatomical image, typically visualized in grayscale; and a Doppler velocity value, which is typically measured in cm / s to measure blood flow velocity in a specific direction and is typically visualized on a red-to-blue color scale.

[0036] Diffusion MRI data

[0037] This modality consists of N-channel 3D imaging data (N>0). Each voxel in the imaging data contains N+1 values. The first value is called the BO signal, and all subsequent values ​​correspond to the diffusion-weighted signal at the voxel location. N is typically 6, but can reach hundreds of channels. This type of modality is often used to explore the intrinsic tissue orientation within organs.

[0038] PET-MRI data

[0039] This mode is generated by a dedicated MRI scanner equipped with PET imaging equipment. It consists of 2-channel 3D imaging data. Each voxel in the imaging data contains two values. The first value corresponds to the MR-weighted signal (which can be T1-, T2-weighted, or any other MR modality), and the second value corresponds to the PET signal. This type of imaging modality is typically used to highlight the concentrated presence of radioactive tracers attached to tumor tissue, superimposed on the structural MRI signal.

[0040] MR (or CT) ultrasound fusion

[0041] While not a modality in itself, 3D or 2D ultrasound data (typically real-time) can be fused with MR or CT data. This can provide a structural / functional view or show features in one modality that might not be so clear in another. This can be used for guidance. The two sets of data can be stored in separate coordinate systems or fused into a single volume, where one modality is registered to the other and then resampled.

[0042] It should be understood that calculations for the masking kernel, opaque channels, and 2D or 3D rendered images can be performed during runtime or can be cached / recorded—particularly in the case of looped (timely) displays, preferably, rendered images are generated and cached during the first loop until the position or kernel parameters are moved. It will also be understood that embodiments of the invention are also applicable to use in real-time image capture scenarios. The user interface can serve as a view replacing that used by a technician to guide the probe while scanning a patient, or as an alternative view for a clinician that can be controlled independently of probe operation.

[0043] The embodiments of the present invention can operate in near real-time, allowing users to navigate the imaging volume and change the content being displayed and not displayed by moving the highlight position and kernel parameters.

[0044] Compared to existing systems that involve slicing through a volumetric plane and then manually cropping the image, it should be understood that embodiments of the present invention offer significant efficiency and flexibility while reducing the expertise and skills required to operate the imaging system.

[0045] The preferred embodiment utilizes full 3D interaction to allow users to select a location in 3D (e.g., via hand tracking or using interactive tools) and to make the structure fade out as it moves away from that point.

[0046] It is understandable that user interactions can be recorded for later replay (and only the viewpoint and parameters used for reproduction need to be recorded, since the view itself can be recalculated during display, which is particularly advantageous if different display devices are used to render 3D image datasets, as different clinicians or experts may have different display technologies available to them). Attached Figure Description

[0047] Embodiments of the invention will now be described by way of example only, with reference to the appended description, wherein: Figure 1 This is a schematic diagram of an imaging system according to an embodiment of the present invention;

[0048] Figure 2a and Figure 2b It is shown that there is no (according to the embodiments of the present invention) Figure 2a ) and there are ( Figure 2bIllustrative line drawings and corresponding images of the processed 3D rendered images;

[0049] Figure 3a and Figure 3b It displays a regular image ( Figure 3a ) and image changes after applying embodiments of the present invention ( Figure 3b Ultrasound scan images; and

[0050] Figure 4 This image shows an image in which the method of the present invention is applied and represents the trade-off between the color distance parameter and the Euclidean distance parameter and the steepness of the Gaussian kernel to Q. Detailed Implementation

[0051] Embodiments of the present invention relate to methods and systems for displaying and applying user input to manipulate 3D images.

[0052] There are many 3D image data sources, including 3D imaging scanners. Embodiments may receive data directly from 3D image data sources, or they may receive data that has been previously acquired and stored in a data repository or the like.

[0053] 3D image data is typically encoded as an array of 3D voxels. In 3D imaging, the term "voxel" refers to a scalar or vector value on a regular grid in three-dimensional space. Like pixels in a bitmap, voxels themselves do not typically explicitly encode their positions (their spatial coordinates) along with their values. Instead, the rendering system infers the voxel's position based on its location relative to other voxels (i.e., its position within the data structure that makes up a single volumetric image).

[0054] In embodiments of the invention, 3D image data is preferably processed (preferably in real-time or near real-time) to suppress image features at the periphery of the field of view. Preferably, the system determines how / whether to depict image features in the rendered output based on a distance-dependent opacity map. In this way, image features at the focal point (specified by the user interface) are displayed in a completely opaque manner, image features around it become less noticeable as the opacity decreases, and image features further away are suppressed more. In one embodiment, the more features are directly within the field of view, the more they are suppressed. It should be noted that the 3D image data is processed as an array of voxels (or something else if voxels are not used). Therefore, the presence of structures is independent of the system and requires no additional processing. Opacity varies depending on the distance from the focal point and color differences (or differences from other scalar values, if not color). Vessels may have similar colors, so the voxels of the vessel will have similar opacity depending on their distance from the viewpoint.

[0055] Figure 1 This is a schematic diagram of an imaging system according to an embodiment of the present invention.

[0056] The imaging system includes an image data source 10, a processor 20, a display 30, and a user interface. In this embodiment, the user interface includes position control 40 and a user input device 45, although it is understood that different display and input devices can be used.

[0057] Processor 20 receives image data from image data source 10 and position data from position controller 40. Processor 20 generates an opacity channel from the position data and uses the opacity channel to render the image data for display on display 30.

[0058] In the illustrated embodiment, the position control 40 is separate from the display 30. In some embodiments, the position control 40 may be superimposed on the image displayed on the display 30. In other embodiments, the position control 40 may be displayed independently.

[0059] The user (who may or may not be the operator of the imaging probe that generates the imaging data provided from the imaging data source 10) interacts with the position controller 40 to define the highlight position (base of arrow (A)) and direction (direction of the arrow). In this embodiment, this is data provided to the processor 20.

[0060] For example, positioning can be accomplished using a mouse, tablet, X / Y / Z position, and the X / Y / Z highlighted direction using a keyboard, slider, etc. In the illustrated example, the position cursor is represented by an arrow and moves from position A to position B.

[0061] Once the location and kernel parameters are determined and the opacity channel is calculated... The generated 2-channel image is then output through a transfer function for visualization. This transfer function maps intensity to color and the calculated opacity channel to opacity.

[0062] Figure 2a and Figure 2b This indicates that according to the embodiments of the present invention, there is no ( Figure 2a ) and there are ( Figure 2b ) processed 3D rendered images, and Figure 3a and Figure 3b These are ultrasound scan images showing changes after applying embodiments of the present invention. Figure 3a These are illustrative line drawings and corresponding images (illustrated illustrations without applying embodiments of the present invention).

[0063] Given the intensity and opacity channels, the application of the transfer function by the processor 20 is straightforward. It can be understood that the output can be directed to a 3D display device, projecting a 3D image onto a 2D display, or the output can be rendering data (or a basic 3D image dataset and opacity channels, or just the opacity channels).

[0064] It's understandable that the target location and kernel parameters... All of these can be adjusted interactively. Preferably, the system includes a user interface in which the user can move a cursor to select a target point and use sliders or other GUI elements to select kernel parameters.

[0065] The amount of surrounding area hidden can be controlled by kernel tradeoff parameters (as described above, preferably provided to the user in the form of a GUI slider, etc.). The tradeoff in the above embodiment is color distance. Between and Euclidean distance, and through The steepness of the opacity kernel, such as Figure 4 As shown.

[0066] For example, Figure 3a This is an example of routine rendering of ultrasound image data used in medical imaging to help clinicians make diagnostic or treatment decisions. Figure 3b This is an image rendered using an embodiment of the present invention. In an embodiment of the present invention, the rendering is preferably modified according to user input, thereby rendering organs or other imaging structures at or around the highlight focus point. Figure 4 (The crosshairs in the image), but when encountering image features far from the highlight focus, these features are suppressed relative to the distance to the highlight focus.

[0067] Preferably, the system includes a user interface that allows the user to interact with the rendered 2D environment in a 3D manner. The user interface allows the user to select a location in 3D (e.g., by manual tracing or using interactive tools) and to make the structure fade away from that point (see...). Figure 2b and Figure 3b ).

[0068] In this preferred embodiment, 3D image data in scalar (1-channel) or vector (multi-channel) image form is used as input. The system calculates the opacity channel based on a kernel that acts on the intensity of a voxel in the 3D image data and the relative position of that voxel with respect to a user-defined location (typically, the system will have a default position that the user can manipulate via a user interface). It is understood that image data in other formats can also be used as input.

[0069] The opacity channel is calculated relative to the highlight focus, and this opacity channel is used to generate the image from the input image data. Figure 2b or Figure 3bThe rendered view. From Figure 2b and Figure 3b As can be seen, the area of ​​interest is opaque, but is visible through the semi-transparent structure.

[0070] Use this transfer function to visualize 3D images, preferably using volume rendering that produces a 2D projection.

[0071] As will be understood, volume rendering refers to a set of techniques used to display a 2D projection of a 3D discrete sampled dataset, typically a 3D scalar field. To render a 2D projection of a 3D image dataset, the camera is defined in space by its volume, opacity, and color relative to each voxel. This is typically defined using an RGBA (for red, green, blue, and alpha) transfer function, which defines an RGBA value for each possible voxel value.

[0072] For example, a volume can be viewed by extracting isosurfaces (surfaces of equal value) from the volume and rendering them as polygonal meshes, or by rendering the volume directly as data blocks. The cubes algorithm is a common technique for extracting isosurfaces from volume data. Ray casting algorithms are a common technique for directly rendering volumes.

[0073] Preferably, the 3D image dataset is stored with a uniform grid. A D-dimensional scalar map of the samples. This can be done as a transformation step at the points received in the 3D image dataset, or alternatively, the dataset can be stored upon receipt and transformed / mapped as an initial step when rendering is to be performed.

[0074] Will Defined as a D-dimensional scalar graph, where the grid... The samples on, then It is a D-dimensional scalar image. Similarly, Defined as a graph with vector values, and It is an image that takes values ​​from a D-dimensional vector. In the following text, we will represent all images as... And assume that the scalar image is a vector image, where d=1.

[0075] To calculate the opacity channel, preferably, the user provides:

[0076] 1) Spatial location (preferably via a movable cursor) as well as;

[0077] 2) Used to mask the kernel An M-dimensional parameter vector (which can be provided, for example, through a slider or other controls in the GUI).

[0078] In one embodiment, masking the kernel Position and images Mapped to a scalar opacity value, and has the following form:

[0079]

[0080] For example, the kernel can use an isotropic Gaussian kernel, to... Centered on:

[0081]

[0082] in It is a scalar value representing the width of the Gaussian kernel.

[0083] From the discussion above, it is clear that the kernel does not necessarily have to be Gaussian. Other examples include radial (spherical) step functions and inverse Gaussian kernels:

[0084] i) Radial step function, Centered on:

[0085]

[0086] in It is a scalar value representing the radius of the radial kernel.

[0087] ii) with A centered inverse Gaussian kernel: (This hides the target region, allowing other regions to be viewed):

[0088]

[0089] in It is a scalar value representing the width of the Gaussian scalar.

[0090] Extending the above method to any kernel, a preferred embodiment uses a kernel that combines intensity (relative to a reference intensity value) and location (Euclidean distance to the target of interest) to define the opacity channel. As shown below:

[0091]

[0092] in Opacity is controlled by intensity. =0) or opacity determined by a location-based kernel ( =1) Trade-off factors between controls, It is a location-based kernel, such as any kernel as described above, and It is a strength-based kernel, for example:

[0093]

[0094] As in the above situation, It is a reference image value (which can be the intensity at the target of interest, or typically fixed in scalar ultrasound). =255, which is the intensity of the bright white area.

[0095] It is understandable that these parameters do not need to be provided by the user and can be system defaults. Furthermore, the location and masking of kernel parameters can be provided by an external system that has already recorded previous views of the dataset or has data from other sources (diagnosis, imaging, medical history, or other data) and highlights potentially interesting features guided by that data. This system can also include machine learning or other systems to assist in the optimal selection of parameters for specific features located at the focal point or highlighted position (e.g., within the crosshairs).

[0096] Figure 4 An example image of an opacity channel obtained using an embodiment of the present invention is shown. The image represents selecting a point on the atrium. The opacity channels (white cross) are increased by the values ​​of Q (from left to right: 0.1, 0.2, 0.3, 0.4, and 0.5) and Q (from top to bottom: 0.05, 0.1, and 0.15).

[0097] It should be understood that the above methods can be implemented in software and / or hardware. A recently developed technique to accelerate traditional volume rendering algorithms—such as ray casting—is the use of modern graphics cards. Starting with programmable pixel shaders, the ability to perform parallel operations on multiple pixels was recognized, leading to the performance of general-purpose computations on graphics processing units (GPGPUs) and other high-performance hardware. Pixel shaders can randomly read from and write to video memory and perform some basic mathematical and logical calculations. These single-instruction multiple-data (SIMD) processors are used to perform general computations such as rendering polygons and signal processing. In recent generations of GPUs, pixel shaders are now capable of functioning as multiple-instruction multiple-data (MIMD) processors (now capable of independent branching), allowing the use of up to 1GB of floating-point format texture memory. With this capability, almost any algorithm with steps that can be executed in parallel can be performed at a tremendous acceleration, such as volume ray casting or tomographic reconstruction. Programmable pixel shaders can be used to simulate variations in properties such as lighting, shadows, reflections, and the color of emitted light. Such simulations can be written using high-level shading languages.

[0098] For illustrative purposes, the foregoing preferred embodiments have been disclosed. Variations and modifications to the basic concepts of the invention will be apparent to those skilled in the art. For example, graphic symbols other than dots or crosshairs can be used to depict positions within a volume. The user interface is not limited to specific software elements. Not only can different software GUI elements be used, but hardware interface features such as trackballs, joysticks, rotary switches, and buttons can also be used. A mouse, joystick, controller, slider, or other input device can also be used when motion-based detectors, virtual controllers / environments, augmented reality, etc., can be employed. It will also be understood that the resulting rendered images can be used with many different display technologies, including 2D, 3D, virtual reality, augmented reality, holography, and other display types. All these variations and modifications are intended to be covered by embodiments of the invention.

[0099] It should be understood that certain embodiments of the invention discussed below can reside as code (e.g., software algorithms or programs) in firmware and / or on a computer-usable medium with control logic capable of execution on a computer system with a computer processor. Such a computer system typically includes a memory storage device configured to provide output from code execution, which configures the processor according to the execution. The code can be arranged as firmware or software and can be organized as a set of modules, such as discrete code modules, function calls, procedure calls, or objects in an object-oriented programming environment. If modules are used to implement the code, the code can comprise a single module or multiple modules that cooperate with each other.

[0100] The alternative embodiments of the present invention may be understood to include the components, elements and features mentioned or indicated herein, individually or collectively, in any one or two or more or all combinations of any of the components, elements or features, and wherein specific integers having known equivalents in the field to which the present invention relates are mentioned herein and are considered to be incorporated herein as if set forth separately.

[0101] Although illustrated embodiments of the invention have been described, it should be understood that various changes, substitutions, and modifications can be made by those skilled in the art without departing from the invention as defined by the claims and their equivalents.

[0102] This work was an independent study funded by the National Institutes of Health (Invention under the Innovation Program, 3D Heart Project, II-LA-0716-20001). The views expressed are those of the authors and not necessarily those of the NHS, NIHR, or the Department of Health.

[0103] This application claims priority to GB 1913832.0, the contents of which, along with the contents of the abstract attached to this application, are incorporated herein by reference.

Claims

1. A computer-implemented method for navigating and displaying 3D image data, comprising: Retrieve the dataset of 3D images to be displayed; Receive the identifiers of the highlighted locations within the 3D image dataset; Calculate a scalar opacity map of the 3D image dataset, the scalar opacity map having a value at each of a plurality of locations in the 3D image dataset, the corresponding value depending on a corresponding distance from the highlighted location, and depending on the value of the 3D image at the corresponding location relative to the value of the 3D image at the highlighted location; as well as Opacity is applied to the 3D image dataset to generate a modified 3D image view. The calculation step uses a masking kernel that maps the location and image to scalar opacity values.

2. The method according to claim 1, wherein, The masking kernel is selected from a set including Gaussian kernels, kernels based on uniform / rectangular distributions, radial basis functions, spherical step functions, or exponential distributions.

3. The method of claim 2, further comprising receiving the highlight locations in the 3D image dataset via a user interface, the method further comprising calculating the scalar opacity map depending on the highlight locations, the 3D image dataset, and the masking kernel.

4. The method of claim 3, further comprising receiving parameters of the masking kernel via the user interface.

5. The method according to claim 3 or 4, wherein, The user interface includes a 2D representation of the 3D image dataset.

6. The method according to claim 1, wherein, The calculation steps use predefined shapes or predefined 3D as masking functions.

7. The method according to any one of the preceding claims further includes rendering the modified 3D image as a 2D or 3D image for display.

8. The method according to any one of the preceding claims, further comprising calculating the scalar opacity map for a region of the 3D image dataset, said region including a portion of the 3D image dataset in the field of view between the highlighted location and the edge of the 3D image dataset.

9. A system for navigating and displaying 3D image data, comprising: A data repository that stores datasets of 3D images to be displayed; The user interface is configured to receive identifiers of highlighted locations within the 3D image dataset; The processor is configured to: calculate a scalar opacity map of the 3D image dataset, the scalar opacity map having a value at each of a plurality of locations in the 3D image dataset, the corresponding value depending on a corresponding distance from the highlighted location and depending on the value of the 3D image at the corresponding location relative to the value of the 3D image at the highlighted location; Furthermore, opacity is applied to the 3D image dataset to generate a modified 3D image view. The calculation uses a masking kernel that maps the location and image to scalar opacity values.

10. The system according to claim 9, wherein, The system is configured to receive adjustable parameters of the masking kernel via the user interface, and the processor is configured to apply the received parameters when calculating the scalar opacity map.

11. The system according to claim 10, wherein, The user interface includes a display showing a 2D representation of the 3D image dataset, and the system is configured to receive a designation of the highlighted position via the 2D displayed representation.

Images