High Definition Volume Rendering is a CPU-based, advanced visualization platform that allows original equipment manufacturers (OEMs) in a variety of fields to deliver unparalleled quality, performance and scalability to their customers. Fovia's flexible, innovative XStream HDVR platform overcomes the inherent limitations of other imaging technologies, and enables local, enterprise-wide and remote volumetric rendering with off-the-shelf hardware. XStream HDVR delivers high fidelity 2D, 3D, 4D and 5D (e.g., fusion 4D) real-time rendering anytime, anywhere. By natively integrating XStream HDVR, OEMs can control their own proprietary product workflows. See Technology to learn more.
Fovia supports all fields that use volumetric data, including Medical, Dental, Oil & Gas, Pharmaceutical, Scientific, Industrial, Seismic, Veterinary and 3D Printing. Visit Markets to learn more.
Fovia’s customers, such as GE Healthcare, IBM, Pfizer, Johnson & Johnson and Schlumberger, are widely respected in their fields. United States agencies such as NASA and the Department of Defense also use High Definition Volume Rendering. For a representative list, see Customers.
Very easy. The component architecture of XStream HDVR provides multiple points of integration for OEM customers, allowing the easy incorporation of example code within proprietary applications. XStream HDVR is available in C++, Java and .NET, with bindings that support zero-download clients through HTML5 and mobile devices. See How to Integrate to learn more.
A Software Development Kit (SDK) is a set of software development tools that enable the creation of applications for a specific software package or development platform. Fovia provides the XStream HDVR SDK Suite for integration via implementation of its APIs and libraries within OEM imaging products.
Why start from scratch and use a great deal of your R&D resources when Fovia has already created a solution that has been recognized as best-in-class? We know of many companies that have spent years developing their own 3D, and have come to us for integration once they realize that Fovia’s technology is superior. Integrating XStream HDVR allows your company to focus on its own product line development, while being able to offer world-class 2D, 3D, and 4D.
It is never too late to switch to a more scalable technology. A small amount of work in the short-term can provide significant long-term benefits. Once an XStream HDVR integration is complete, all volumetric data can be rendered on the CPU, including polygon objects, texture maps, UI components and any other renderable objects.
Fovia’s XStream HDVR engine provides anytime, anywhere image access. The option to invoke an existing GPU engine in executable code is available, this may limit where your software will run.
How does XStream HDVR perform in the cloud or in large-scale enterprise deployments compared with GPU-based solutions?
The XStream HDVR CPU-based solution is superior to GPU solutions for many reasons.
For the IT staff, deployments are greatly simplified. The server can be deployed across the enterprise or cloud since it can be virtualized. Additionally, XStream HDVR has no client hardware requirements because viewing is via a thin-client or a zero-download HTML5 viewer.
CPU-based rendering eliminates the challenges imposed by GPU solutions, giving the end user a richer experience. Interactive and final quality are virtually identical because there is no downsampling. The developer is freed from the complexities of GPU architecture and the hardship of compatibility testing (e.g., an NxM matrix of video cards vs. different versions of Windows).
XStream HDVR offers many additional benefits over GPU solutions including the ability to render large datasets with uncompromised interactive quality and performance, deliver high fidelity image quality (even on the largest monitors), run clients anywhere without compatibility issues (on any legacy hardware), and more—just ask or visit CPU vs. GPU!
XStream HDVR offers the significant development benefits over a GPU implementation:
- GPU programming is more difficult than CPU programming. XStream HDVR deploys smart algorithms, with clear advantages and much more flexibility than the “brute force” approach required by GPUs and OpenGL.
- XStream HDVR eliminates the GPU quality assurance headache—MxN matrix of each flavor of Windows against each type of video card—and yes, each one has to be tested!
- XStream HDVR allows integrators to build more flexible and less complex long-term implementations that are easier to maintain.
In addition, Fovia’s CPU-based solution deploys easily across the enterprise or to the cloud, lowers equipment cost, needs no special purpose hardware or graphics cards, requires less real estate, less power, and therefore produces less heat, and facilitates deployment to any number of clients.
Many times, developers assume that a GPU-based solution is “easier to implement,” and proceed down this path without doing a complete analysis. We do our best to educate potential customers about the pros and cons of various solutions so that they are able to make well-informed decisions. See CPU vs. GPU to learn more.
While that may be true, each customer determines how to use XStream HDVR to differentiate itself from its competitors. With XStream HDVR, Fovia’s OEM customers can enhance their own proprietary workflows and user interfaces, or develop new ones based on XStream HDVR. XStream HDVR allows High Definition Volume Rendering to be natively integrated into Fovia’s customers’ applications. Unlike many 3D solutions that are “stand-alone” plug-ins requiring end users to toggle back and forth between third-party developed pages, XStream HDVR enables OEM-driven, proprietary, embedded solutions.
For example, our many PACS customers’ implementations look very different from one another, even though they are all built using XStream HDVR.
Does any other company deliver the quality and performance of the XStream HDVR engine offered by Fovia?
No. It is too computationally complex for traditional GPU-based volume rendering methods to achieve interactive High Definition Volume Rendering, and other CPU-based solutions cannot match the quality and performance of our software, developed and continually improved during the past 10+ years. Visit Quality & Performance.
It’s real and is a direct result of deep supersampling and our ability to highlight and independently adjust different aspects of the data, such as bone, muscle or vessels. While this may sound complicated, XStream HDVR is able to produce these images interactively and in real-time. Visit Quality & Performance.
See it for yourself. A side-by-side comparison is the best way to appreciate High Definition Volume Rendering. We strongly encourage potential customers to conduct a trial rendering of the same, self-selected datasets using both XStream HDVR and any other volume rendering engine of their choice (either their own or one of our competitors’). Compare the quality and performance side-by-side. We’re confident that once you experience High Definition Volume Rendering, you will want to switch to XStream HDVR. Contact Us to inquire about arranging a side-by-side comparison.
Visit Quality & Performance to learn more.
No. Although most of our competitors use these techniques, the superior quality of High Definition Volume Rendering would not be possible if we did. The supersampling algorithms of XStream HDVR make its high quality interactive rendering possible. Request a demo to see our interactive supersampling for yourself!
No. All volume rendering using XStream HDVR is done in realtime.
Other volume rendering solutions claim they can handle thousands of slices, but this is not the same as being able to render them. These solutions typically downsample or resample sub-volumes of the original data, as they are unable to truly render the entire dataset. Our XStream HDVR engine can actually render these large volumes, resulting in superior interactive performance and quality.
Fovia talks about the “data explosion.” At a certain point, won’t the dataset size become too big to render in real time?
No. Our solution’s performance has increased by over 100x since we developed our first solution. Fortunately, the processing power of CPUs has been following Moore’s Law for decades, doubling performance approximately every 18 months. While the “data explosion” is real, it has not kept pace with CPU performance growth. Since XStream HDVR is purely CPU-based, it is extremely scalable, and actually outpaces dataset growth. In addition, the performance gap between XStream HDVR and other solutions increases as dataset sizes increases.
Yes. Fovia’s XStream HDVR engine works with general purpose, off-the-shelf hardware and does not require a GPU.
No. Both of these approaches have significant scalability limitations. XStream HDVR is a true software-only solution.
No. XStream HDVR does not require OpenGL, DirectX or any third-party 3D graphic library. Our code is self-contained.
On the server side, Microsoft Windows, Macintosh OS X and Linux (both Ubuntu and Red Hat distributions) are supported. On the client side, Microsoft Windows, Macintosh OS X, Linux and any web-enabled device are supported.
Flexibility is critical in today’s market, as user requirements can be complex and often change rapidly. Fovia provides the broadest and most flexible platform options in the industry, with multiple client-server configuration options available for maximum platform interoperability.
Fovia’s OEM customers can easily and seamlessly mix and match their client-server configurations with Mac, Windows, Linux, mobile, cloud and web platforms.
Visit F.A.S.T. Cloud SDK to learn more.
Are there any performance differences between your F.A.S.T. Cloud SDK implementation and your other numerous implementations?
When compared with GPU-based volume rendering solutions, XStream HDVR is able to address an order of magnitude more data without compromising image quality. Other CPU-based solutions need significantly more memory to even approach the quality and speed of XStream HDVR, and still cannot match its high performance.
Specifically, XStream HDVR requires that the dataset reside in memory, using only ~40% memory above the dataset size (+/- 5% depending upon dataset characteristics). For example, a 100MiB dataset on disk (two bytes per voxel) will use about 140MiB of total memory.
What is the minimum recommended bandwidth? How does XStream HDVR perform at lower bandwidths, such as a 1 Mbps connection?
Fovia recommends 4 Mbps bandwidth. However, XStream HDVR can certainly run well at 1 Mbps and 100 ms ping time, providing an interactive experience of 10 frames/second (close to theoretical best case). In certain circumstances, it may be acceptable to run XStream HDVR using bandwidth as slow as 768 kb and 200 ms pings.
As with any volume rendering solution, bandwidth certainly affects end user interaction, but by itself does not affect rendering time on the server. XStream HDVR achieves excellent quality and performance even in low bandwidth situations (e.g., 1 Mbps and 100 ms ping time). Furthermore, Fovia’s API allows users to interactively adjust quality and performance to meet their specific needs.
The answer to this question depends upon many variables. Unlike other systems, XStream HDVR does not require duplicate copies of the data, but instead performs processing on the original, single copy of the dataset. Additionally, XStream HDVR requires minimal memory overhead‑only about ~40% of the original data size—whereas many solutions require memory overhead 2x-4x the original data size. Contact Us to discuss your specific implementation.
Because Fovia doesn’t require any pre-processing like many other solutions, XStream HDVR loads data almost instantaneously from your application’s memory space or disk. In most cases, the time-to-first-image will be less than one second because the original image data (DICOM, SEG-Y, RAW or other cross-sectional data) will already have been loaded into your application’s memory space.
Since XStream HDVR rendering is done on the server, as long as the server hardware is sufficient, Fovia’s client-server based architecture allows otherwise outdated desktop computers and laptops to be used as fully-functional clients. Another solution to extend the useful life of your hardware is our zero-download client available via our F.A.S.T. Cloud SDK.
While I understand the benefits of client-server rendering, my device sits in the operating room and must be self-contained. Is XStream HDVR an option for me?
Yes. Both client and server can certainly run on the same system, so that the system is self-contained. This is one deployment strategy used by some of our customers. A key benefit of XStream HDVR architecture is that with a simple preference change your application could, if you choose, also support client-server in the future without any performance impact.
Fovia provides a DICOM data loader, and the XStream HDVR server API provides an interface that allows customers to implement their own DICOM loader, if desired.
XStream HDVR supports parallel and perspective volume rendering, MIP, Faded MIP, MPR, Curved MPR, RaySum and MinIP. XStream HDVR also supports 2D, 3D, 4D rendering, MPR oblique and double oblique views, MPR cross-sectional views, fusion and more—see our XStream HDVR Feature Highlights.
XStream HDVR supports multi-modality fusion of multiple datasets, including MPR and 3D fusion (to provide landmarks in a volumetric view). Cross-sectional modalities can also be registered, such as CT, MR, PET, SPECT and in certain cases, live ultrasound. View sample images in our Gallery.
Yes. This is an essential element of Fovia’s solution. The XStream HDVR engine gives users maximum flexibility to modify ALL rendering parameters in real time, without preprocessing, and without any speed or quality penalties.
Fovia provides two general tools—cut planes and free-hand cut (also know as virtual scalpel). In addition, Fovia provides more sophisticated tools, such as min/max threshold, spectral analysis, and F.A.S.T.® Interactive Segmentation. See F.A.S.T Interactive Segmentation to learn more.
The XStream HDVR engine is highly optimized to incur minimal rendering slowdown during static and interactive segmentations. Using the XStream HDVR SDK Suite, OEMs can create custom segmentations unique to their workflows and specific applications. See Product Overview.
F.A.S.T. Interactive Segmentation provides dynamic, real-time, interactive 3D segmentation functionality. Developers can quickly create rich F.A.S.T. Interactive Segmentation applications, powered by Fovia’s high performance XStream HDVR.
Continuous control of segmentation parameters and access to instantaneous results via high fidelity 3D rendering enable accurate segmentation, adjustment and analysis in seconds. Users can quickly and easily segment and modify “on-the-fly” anatomic regions of interest for quantitative imaging, radiation therapy, surgical planning, robotic surgery, 3D Printing and more!
Additionally, F.A.S.T. Interactive Segmentation is designed to complement existing, automatic segmentation algorithms, allowing rapid, real-time “clean-up” of data to meet radiologists’ exacting requirements.
Whether used alone or in conjunction with other segmentation algorithms, F.A.S.T. Interactive Segmentation allows physicians to reduce complex case times from hours to minutes, thereby expediting reporting and streamlining diagnoses. See F.A.S.T. Interactive Segmentation for more.
Yes. XStream HDVR supports linear and volumetric point-and-click measurement that is accurate to the subvoxel level.
Yes. XStream HDVR offers the ability to fuse multiple datasets, easily correlate navigation systems to the data, and visualize all synchronizations in volume rendering and MPR views, thereby providing advantages in robotically-assisted minimally invasive surgeries. In addition, the XStream HDVR SDK Suite makes this functionality available in stereo.
Yes. F.A.S.T. RapidPrint provides the ability to export a watertight polygon mesh of the entire volume or a segmented region to a variety of data formats including STL, PLY and OBJ. The benefit of this approach is that editing, cleanup and preview can happen within the user’s application, avoiding the need for a third party application for these functions. XStream HDVR also supports one-click printing. See F.A.S.T. RapidPrint to learn more.
Without using a GPU, XStream HDVR supports real-time rendering and direct fusion of semi-transparent polygonal and volumetric data. Texture map rendering of polygon data from image files, including facial textures and live render-to-texture sources, is also supported. In addition, exported polygonal regions that have direct volume-to-polygon color mapping can be viewed and updated in realtime within the original volume.
Yes. The XStream HDVR SDK Suite supports importing of OBJ and STL files. Real-time editing of volume data and exporting to a variety of formats, including OBJ, PLY and STL is also supported. Additionally, Polygon-based geometry can be dynamically constructed with the XStream HDVR SDK Suite.
Yes. XStream HDVR supports CAD/CAM and rapid prototyping for both design and manufacturing.
Yes. XStream HDVR offers fully interactive, real-time 4D manipulation (including gated cardiac studies). Visit our Movie Gallery for an example of 4D rendering.
Fovia’s solution supports output images up to 8192 x 8192 pixels. Supported image formats include 24-bit RGB, 256 levels of gray, and JPEG. XStream HDVR movies are output in Quicktime file format, up to 8192 x 8192 pixels.
We created the movies with the XStream HDVR SDK Suite, capturing individual key frames that contain a static JPEG image and the rendering state of the frame. The movie files, a set of key frames and timing transitions between frames, were saved as a script. Next, the rendering engine interpolated the details between two frames to produce the animation sequence.
XStream HDVR can play this script back in real time, or generate high resolution movies for playback.
Yes. The ability to deliver non-compromised remote renderings is one of the main advantages of Fovia’s software. With XStream HDVR architecture, any networked desktop, laptop or mobile device can serve as an advanced 3D workstation.
The non-video-card-based architecture of XStream HDVR provides near-linear scalability for up to 256 cores/CPUs. This is a distinct advantage over solutions that rely on “brute force” volume rendering and/or video cards that may saturate memory bandwidth with just one or two cores/CPUs. Visit Scalability to learn more.
Extremely scalable. To illustrate, let’s compare two datasets—512 x 512 x 512 vs. 512 x 512 x 4096. The 4096-slice dataset is 8 times larger than the 512-slice dataset, and one might expect it to be rendered 8x more slowly. With our XStream HDVR engine, however, the speed difference is not 8x, but approximately 10%. Visit Scalability to learn more.
Extremely scalable. Unlike many alternative approaches, XStream HDVR does not stretch images, and all images—large or small—are generated directly by the XStream HDVR engine. To illustrate scalability, let’s compare two projection planes—512 x 512 vs. 1024 x 1024. The larger projection plane is 4 times bigger than the smaller one, and one might expect it to be rendered 4 times more slowly. With Fovia’s XStream HDVR engine, however, the speed difference is not 4x, but approximately 1.7x. Visit Scalability to learn more.
At Fovia, we believe that scalability is one of the most important issues facing the advanced imaging industry. The sheer volume of data being acquired by modern scanners continues to grow rapidly, placing increasing pressure on volume rendering technologies. Fovia’s solution is designed with algorithms and architecture that take full advantage of future advances in both medical imaging and computing—larger datasets, larger projection displays, multi-core processors, multi-CPU environments and clustering—without sacrificing quality or performance. Not only does the XStream HDVR engine provide unparalleled image quality and rendering performance today, but it is designed to anticipate and capitalize on future technological advances. This will broaden the gap between Fovia and its competitors, as solutions that rely on specialized hardware and/or video cards are at a significant disadvantage in terms of scalability.
Yes. XStream HDVR is ideally suited for web deployment and ease of integration.
XStream HDVR excels in a virtualized environment, capitalizing on its CPU-based architecture and ability to scale with the number of cores. Implementation is greatly simplified without the complications typically encountered with GPU deployments in the cloud, while still maintaining Fovia’s award-winning speed and performance.
The amount of data needed to represent the XStream HDVR engine state of any rendering window is very small, approximately 4-5K. This can be replicated easily across all nodes to ensure that upon any network loss, its state can be reconstructed quickly by another node and served up to the original client.
This depends on a variety of factors including the size of the data, the workflow, and the type of rendering. (For example MPR mode requires the least processing power.) The limiting factor is generally how many datasets can be fit into memory. XStream HDVR requires minimal memory overhead—only about ~40% of the original data size—whereas many solutions require memory overhead 2x-4x the original data size. XStream HDVR performs well in very large deployments (Visit Customers to see examples of companies currently utilizing XStream HDVR for their offering). Contact Us to discuss your specific implementation.
See Scalability to learn more!
No. XStream HDVR utilizes all available cores and runs at the lowest priority level to ensure that the user experience is never compromised. The rendering engine automatically adjusts the rendering quality based upon the available processing power and evenly distributes processing loads across all cores. The XStream HDVR SDK Suite also supports a “Desired FPS (frames per second)” feature that instructs the rendering engine to automatically adjust image quality to achieve a desired frame rate.
As few as four function calls are needed to initialize the XStream HDVR library, load a dataset and render an image into a window.
Sample applications are provided to demonstrate XStream HDVR integration with a range of GUI platforms. The XStream HDVR SDK Suite also includes platform-independent development classes that greatly simplify integration with mouse and multi-gesture touch adaptors. More important, the API does not assume deep knowledge of volume rendering or 3D graphics. Fovia will take care of 3D volume rendering so that you can focus on your application!
I have my own custom segmentation and data analysis algorithms. Can these be used with the XStream HDVR SDK Suite?
Yes. Developers using the server plug-in API can incorporate their own segmentation and data analysis algorithms into the XStream HDVR server.
Fovia’s technical staff is available to answer any development or technical questions that may arise. In addition to our Documentation, we can also provide additional sample code or arrange screen-sharing sessions to address your questions.
Fovia’s technology is licensed on an OEM basis to modality manufacturers, PACS companies and other companies that can benefit from integrating our High Definition Volume Rendering® into their products. If you represent an OEM interested in integrating our solution, visit Get XStream HDVR to request a demonstration.
End users (e.g., physicians) who want to use our solution can tell their advanced visualization vendors that they need High Definition Volume Rendering by Fovia, or Contact Us and we will follow up with them.
Quite simply, our success depends on the success of our customers. Fovia is committed to providing our customers with the very best technological solution and a business model tailored to their specific needs. Contact Us to request a quote.
Visit Get XStream HDVR to request a trial version.