CUDA Spotlight: Ismayil Guracar

Ismayil Guracar

Real-Time Signal Processing in High-Performance Diagnostic Ultrasound Imaging

This week's Spotlight is on Ismayil Guracar, a Senior Key Expert at Siemens Medical Solutions USA.

Ismayil and his colleagues have applied CUDA to the high-data-rate signal processing pipeline of the ACUSON SC2000™ ultrasound imaging system.

In addition to replacing existing FPGA- and CPU-based processing functions, the team has been able to rapidly extend the capabilities of the system to achieve improved speckle reduction, information content and lesion conspicuity.

Q & A with Ismayil Guracar

NVIDIA: Ismayil, tell us about your work at Siemens Medical Solutions.
Ismayil: I work in the Innovations Applications Group in the Ultrasound Business Unit developing new ideas. I manage external university collaborations in the areas of ultrasound contrast and molecular imaging. I also work with internal developers to advance and deploy signal processing algorithms for diagnostic medical ultrasound.

I have been an enthusiastic advocate of GPUs in our organization and have been teaching and leading discussion groups on CUDA programming for several years.

NVIDIA: What are some of the key challenges in the field of ultrasound?
Ismayil: Ultrasound is a real-time imaging modality with a lot of interesting signal processing problems. Because ultrasound is a coherent imaging modality, speckle noise is an important factor which limits lesion conspicuity. There are a number of techniques to reduce speckle that rely on sophisticated signal processing. These techniques happen to map very nicely to GPUs.

Example of a thyroid image, without and with spatial compounding speckle reduction
Example of a thyroid image, without and with spatial compounding speckle reduction.

Accurate assessment of dynamic heart function is a critical need. Real-time 3D ultrasound imaging of the heart (4D) requires processing at least 100 megasamples per second with minimal latency, which is a big challenge

Collecting all of the information needed for an ultrasound exam in a single heartbeat without stitching together data from multiple heartbeats enables better quality imaging and speeds up example time. This is especially true in many disease states with irregular heartbeats and also enables real time interventional guidance and automatic measurements.

GPU computing enables us to apply signal processing algorithms to dramatically improve real-time ultrasound imaging of the heart.

Volume cardiac image with multiplanar reformatting
Volume cardiac image with multiplanar reformatting.

NVIDIA: What role does GPU acceleration play?
Ismayil: We currently use GPUs in the real-time signal processing pipeline of our medical ultrasound instruments for image reconstruction and 3D rendering and for speckle noise reduction signal processing algorithms.

NVIDIA: In what specific ways do you leverage CUDA?
Ismayil: We have used CUDA to replace FPGA and custom ASIC hardware in our signal processing pipeline. We have also used CUDA to very rapidly develop and deploy new algorithms. The FPGA place and route cycle took half a day. With CUDA, compilation takes just seconds. We can go from idea to prototype to product a lot faster.

NVIDIA: What most excites you about the future of ultrasound?
Ismayil: There is still a lot of room for improvement. We are not yet making use of all of the information available and ultrasound is an area where the application of more compute power could make a big difference.

Doing this in a power-efficient (especially for handheld applications) and cost-effective way is a big challenge and I am excited that NVIDIA is investing so much in developing the GPU platform and tools.

NVIDIA: Twenty years from now, what can the average person expect in terms of an ultrasound experience?
Ismayil: Ultrasound is finding its way into more clinical applications, enabled by miniaturization and high-performance, low-power signal processing. We will see a lot more 3D/4D imaging and automatic optimization, making it easier to acquire the needed information.

Over the next decade I also see ultrasound in a larger role in disease screening, diagnosis and treatment with the growth of new parametric imaging modalities such as targeted contrast and elasticity. Ultrasound can provide information on biological processes occurring on a molecular level.

NVIDIA: How did you become interested in this field?
Ismayil; When I saw my first real time color Doppler blood flow image of the heart I got hooked. I was fresh out of grad school and I found this incredibly compelling--I could take things I learned and apply them to understanding and developing life-saving technology.

Since then I have gone back to learn more, taking additional courses in medical imaging and biology and working with others to advance the field. There are so many interesting problems to be solved and there is still a lot of work to be done.

Bio for Ismayil Guracar

Ismayil Guracar has been working in the ultrasound imaging field for over 26 years. He is currently a Senior Key Expert with the Innovations Applications Group at Siemens Medical Solutions USA, Ultrasound Business Unit, in Mountain View, California. He is a graduate of the University of the Pacific (BSEE) and Stanford University (MSEE).

His interests include ultrasound image formation and high-performance real-time signal processing, especially using GPUs. He holds 63 US patents, was a Siemens Inventor of the Year (2002) and has pioneered new ultrasound technologies in the areas of parametric and molecular imaging and contributed to the development of many successful diagnostic medical imaging products.

He is married, with two teenage daughters. When he is not thinking about ultrasound and CUDA, he enjoys cooking, helping teach math to high school students, photography, hiking and exploring the ruins of ancient civilizations, especially in ancient Lycia in Southwestern Turkey. He wonders about the lives of people who created technological and social solutions to cope with the challenges of their environment.

Relevant Links

Contact Info
ismayil.guracar (at) siemens (dot) com