Professional Documents
Culture Documents
I High performance, embedded computers ronment on your hands. To remain competi- including cancer and heart disorders. Some
are having a huge impact in a number of in- tive, OEMS must continuously invest in new applications include real-time imaging used to
dustries. One of the most demanding applica- technologies that can improve performance and provide feedback while performing procedures
tions is medical imaging. As health care also lower costs. But it doesn’t end after the sale. such as positioning a stent or other device inside
providers are under pressure to more accurately For these manufacturers, they need to offer someone’s heart. In other cases, multiple modal-
diagnose more patients, this burden is passed long-life support once the system is in the field. ities (CT, MRI, ultrasound, etc.) are used to
along to OEMs developing next-generation These products are typically big ticket items and provide optimal images. The reason being there
medical imaging equipment such as MRI, CT, can cost well over one million Euros. As you can is no single technique that is optimal for imag-
Digital x-ray and Ultrasound devices. Today, imagine, health care organizations expect long ing all types of tissue. In these cases, compute
health care professionals are demanding lower service life from these systems, and want the systems are use to combine images from each
cost systems that offer improved imaging ca- ability to upgrade as technology evolves. modality into a composite image that provides
pability. These systems incorporate 3-D and, in more information than just looking at the
some cases, 4-D images that help users accu- “Using proven embedded systems in their de- images from different modalities separately.
rately diagnose and treat a variety of conditions. signs can give OEMs an easy, cost effective
method to reap the performance and cost ben- In addition to requiring the best images possi-
Increased image processing complexity, the hall- efits of the latest computing technologies,” said ble, safety is a crucial consideration. Many
mark of this medical equipment, has an enor- Bob Ghaffari with Intel. “Embedded systems imaging systems use x-rays to provide a view of
mous impact on compute requirements. Here we offer additional value over compute solutions what is beneath a patient’s skin. X-ray radiation
will discuss how incorporating the latest em- intended for PC’s or IT servers, in terms of levels must be kept at a minimum to protect
bedded compute modules into imaging systems product stability over many years, specialized both patients and staff. As a result, raw image
can provide users with higher quality images and package size, lower noise, and higher reliability.” data can be extremely noisy. In order to provide
help OEMs reduce design cycle times and lower clear images, algorithms designed to reduce
development and production costs. noise are used to process the raw data and ex-
Multiple Modalities, Oh My! tract the image data while eliminating the
Health care professionals rely on computer-gen- noise. In video imaging applications, data often
Competition is Fierce erated images to illustrate important informa- has to be processed at rates of 30 images per
Today a handful of OEMs supply the majority tion about a patient’s health. High-quality second or more. Filtering noisy input data and
of medical imaging equipment and among images are crucial to accurately diagnose a pa- delivering clear, high-resolution images at these
these few companies competition is intense. tient or determine treatment. As an example, the rates requires tremendous compute power. The
Add to this atmosphere significant erosion in data available in these images is invaluable in the only way to get improved results is to increase
product pricing and you have a frenzied envi- early diagnosis and management of diseases the processing power of each machine.
October 2007 10
MEDICAL IMAGING
Parallel Computing to the Rescue example, the chart below illustrate processing ed software and multi-core servers means larg-
speed improving as the number of cores used er workloads can be handled with fewer ma-
Since a single general purpose processor is increases. Here an imaging edge-detection chines. This leads to increased performance at
unable to provide the necessary processing algorithm that has been modified for parallel a lower cost.
power for high-speed image processing, there is processing is running on a quad-core, dual-
a shift underway to move from serial to paral- processor server. The eight cores available to
lel processing. This divide and conquer method the application deliver dramatically improved
Co-Processors Equally Important
means multiple tasks can be performed in par- performance as the number of threads sup- The latest CPUs offer up to four cores per chip.
allel. This approach has been well known in the ported by the application approaches the There is another class of chips that offer “scores
area of “Super-Computing” for many years, but number of cores. of cores” per chip. Hardware co-processors such
is relatively new to medical imaging systems. as Graphics Processing Units (GPUs) and
As demonstrated in the data below, software Field-Programmable Gate Arrays (FPGAs),
Shifting the focus from increasing linear must be multi-threaded to take full advantage among others, are being used to take paral-
execution (clock) speed to increasing execu- of multi-core processors. Threading is a com- lelization to the next level and deliver added
tions that can be performed in parallel, allows plex undertaking and requires the right tools to performance in embedded systems. Attached to
more data streams to be processed simultane- identify the sections of an application that are the motherboard, they offer supplemental,
ously. Multi-core processors can facilitate well-suited for parallel processing. However, the somewhat specialized processing functionality
many data streams running in parallel. As an payoff is that the combination of multi-thread- that helps offload some of the processing tasks
that the processor would normally handle.
While co-processors aren’t new, advances in
45
line 1 standard graphics hardware, in conjunction
with new programming models and other
40
tools have increased their popularity and they
are becoming more widely used in embedded
35 systems. Today they can play a key role in high-
performance imaging systems.
30
With availability of additional variables (multi-
Time
Processing speed is improving as the number of cores used increases. Several advancements, both hardware and soft-
ware related, have lead to revolutionary changes
in the world of high-performance computing.
These improvements help deliver breakthrough
performance, increased power efficiency and cost
advantages that support larger workloads using
fewer servers. These benefits are amplified when
referring to high-end imaging applications.
11 October 2007
MEDICAL IMAGING
October 2007 12