Achieving 3D Visualization for Chairside Dental Assistance Devices with Embedded APUs
Like many technologies in the healthcare domain, the dental assistance device market is moving quickly to embrace 3D visualization at the point of care via a new generation of video- and graphics-optimized touchscreen panels that can be attached to dental chairs for use by dentists and patients alike. Dental practices are beginning to transition away from conventional 2D x-ray film and light-box illuminators to sleek, chair-side monitors that provide 360 degree image visualization and other advanced graphics-driven capabilities in HD resolution with intuitive multi-touch interactivity a la tablet computers. As this technology matures, it is very possible that devices of this kind could become ubiquitous in healthcare facilities spanning far beyond dentistry.
In the dental domain, these devices equip dentists to assess patients’ dental imagery with much greater accuracy and process efficiency while simultaneously providing new levels of visual detail to their patients—where previously patients struggled to understand diagnoses and treatment recommendations made on the basis of static 2D renderings, they now have a much clearer view of the treatment area and care methodology. When paired with integrated, hand-held intraoral cameras, these devices can provide real-time dental visibility far beyond the coin-sized, one-way reflection of a dentist’s mirror.
In day to day operation, dental assistant devices equipped for multimedia and 3D visualization can enhance care and streamline processes at every stage of the patient visit. Beginning with pre-session hygiene mode, the device can provide onscreen guidance that prompts the dental assistant through legally-mandated water flushing procedures that help prevent bacteria aggregation in the system. During the patient session mode, dentists can utilize a USB-attached intraoral camera—which includes a supplemental foot-switch for hygienic, hands-free operation—and access historical patient data in real-time, swiping, rotating and zooming the touchscreen interface to view diagnostics data and manipulate 2D images and 3D CT scans with ease. During session downtime, the system can display entertainment programming, educational information or advertising content at the dentist’s and/or patient’s discretion.
As you would expect, these capabilities require a very high level of processing performance, particularly the high-speed 3D graphics rendering, with tight hardware/software integration to ensure a high quality user experience. The underlying processing platform must strike an optimal balance of processing performance, 3D acceleration, interoperability with back-office IT infrastructure and patient management systems, and power and cooling efficiency. For White Lion Technologies’ Vision U visual dental assistant devices, currently available with Ultradent dental chairs, this balance was achieved using AMD Embedded G-Series accelerated processing units (APUs).
Architected for Acceleration
While lower-performing, mobile-optimized processors can be adequate for some entry-level devices in this product category, the graphics and CPU performance provided by these processors is far short of what’s required to power high-speed 3D visualization. Meanwhile, many high-end desktop chipsets are of limited suitability for the massively parallel processing required for 3D visualization-driven medical applications, and can introduce significant power and cooling challenges. Embedded APUs, on the other hand, can provide an optimal platform for high-speed parallel processing, both in terms of the APU architecture and supported development tools.
From an architectural standpoint, the single-chip combination of CPU and general purpose GPU (GPGPU) on a single APU maximizes parallel processing throughput. Conventional graphics-integrated chipsets rely on the CPU to interface with the GPU via a North Bridge connection, sending calls to the GPU to invoke code running on the co-processor that then sends results back to the CPU. This serial data processing approach adds considerable memory latency, consumes system power, and sacrifices board space.
With the APU architecture, the CPU is tasked with scalar processing including storage, networking, and memory processing, while simultaneously running the operating system, applications and user interface. Meanwhile the on-die GPGPU offloads graphics and multimedia processing using SIMD parallel processing. Data parallel processing can be offloaded from the CPU to the GPU, freeing up the CPU for compute, memory and I/O requests. This optimized data path, further boosted via shared access to the contiguous memory controller, reduces processing latency and helps to improve real-time graphics processing performance.
OpenGL, the cross-platform open API for hardware-accelerated rendering of 2D and 3D computer graphics, is another key enabler for achieving 3D visualization in dental assistance devices with APUs. OpenGL is well suited for the visual representation of digital volume tomography (DVT) data, which is created as a DVT device records x-ray images via an x-ray tube that rotates around the patient. A volumetric dataset is then generated from the individual images of the cone-shaped x-ray beams during the cone beam reconstruction process, and stored for later display. OpenGL introduces the ability to run sophisticated, massively parallelized algorithms to reconstruct these images for high-speed visualization, representing high-resolution volumetric data in a virtual model of the patient’s teeth and jaw structure.
Low Power, Passive Cooling and Ruggedization
Power and cooling efficiencies are of course critical considerations in the selection of a processing platform for any medical device. In the case of AMD G-Series APUs, average power consumption is as low as 2.3 W1 (for the AMD G-T16R APU) and thermal design power (TDP) profiles range from 4.5 W to 18 W—low enough to accommodate a low voltage 24 V power supply, which is important for electrical safety and is easier to certify. In contrast, a comparably performing heterogeneous CPU+GPU chipset can require a power supply from 110 to 220 V, and will therefore require galvanic isolation in the power chain, which adds design complexity.
The power efficiency and thermal dissipation profiles enabled with embedded APUs can also enable a visual dental assistance device to be passively cooled, which yields huge advantages over fan-cooled designs in dental environments. In these environments, it’s commonly understood that anything that moves will need to be cleaned—for hygienic reasons—and therefore the elimination of a mechanical fan is beneficial. A passively cooled device can also allow for a sealed/ventless enclosure. In this way, visual dental assistance devices can approximate the highly safety-regulated display panels used in industrial HMI applications, eliminating fan-related shock and vibration risks and preventing the ingress of particulates and debris. Ruggedized as if for a factory floor, these display panels can ably withstand the environmental challenges of a dentist’s office.
Size and Synergy
The silicon-level integration of a low-power CPU and discrete-class GPU onto a single APU naturally reduces board space demands and eliminates the need for bulky add-on graphics cards. Factoring in the elimination of the mechanical fan, considerable system size reductions can be achieved with low-power APUs. The slim-profile Vision U visual dental assistance system, for example, spans a 21.5 inch HD screen, and the computer module is embedded within the rear of the display. There needn’t be any cords to or from an adjacent desktop peripheral, which helps to ensure a clutter-free medical environment.
x86 APUs also enable smooth interoperability with back-office IT infrastructure and patient record databases, and introduce additional benefits for applications including remote system maintenance and administration utilizing standard networking protocols. The inherent PC and Internet compatibility afforded, coupled with the wide ecosystem of industry-standard, x86-optimized software, operating systems and development tools can unlock significant efficiencies from product design to product operation.
Balancing the myriad design and processor selection considerations that will shape the development of a 3D visualization-capable dental assistance device can be difficult. With embedded APUs, system designers are afforded an elegant x86 processing platform that can provide high-performance, OpenGL-accelerated 3D graphics processing at a low power profile within a compact footprint. While the era of dental chair-side 3D visualization is close at hand, continued development in this product domain could have a profound impact on the entire medical community—from mammography to radiology and beyond.
1 While running a Winbench 99 business graphics benchmark the AMD G-T16R APU consumed an average of 2.284W. I/O Controller Hub power is estimated based on the measured average power drawn by the I/OCH of .965W during a run of 3DMark 06. System Configuration: AMD G-T16R APU (DVT) at 30°C, “Inagua” Development Board, 4GB 1.35V DDR3, Windows 7 Ultimate. EMB-26