Cutting through the noise with Singanallur Venkatakrishnan

Cutting through the noise with Singanallur Venkatakrishnan

Singanallur “Venkat” Venkatakrishnan is a Wigner Fellow in the Imaging, Signals, and Machine Learning Group at ORNL.
Singanallur “Venkat” Venkatakrishnan is a Wigner Fellow in the Imaging, Signals, and Machine Learning Group at ORNL. (hi-res image)

Media Contact

Stephanie G. Seay, Communications
seaysg@ornl.gov, 865.576.9894

August 3, 2018 — Singanallur “Venkat” Venkatakrishnan is helping scientists get a better view of objects under study by some of Oak Ridge National Laboratory’s most powerful instruments by creating algorithms that turn data into 3D renderings with fewer images. The result is a better understanding of the inner workings of everything from new materials to human protein receptors.

“My work addresses the big challenge of giving scientists the most detailed look possible at objects so they can better understand, for instance, how materials behave at microscopic length scales,” said Venkat, a Wigner Fellow in the Electrical and Electronics Systems Research Division.

Venkat employs a method known as model-based iterative reconstruction (MBIR), which uses advanced algorithms to account for unwanted elements, or “noise” in datasets to produce an accurate 3D model. Over the past decade MBIR has been used by researchers as a way to reduce the amount of radiation patients were exposed to in CT scanners. “We want to be able to look at objects in 3D using neutrons much like what a CT scanner does with X-rays when we look inside a body,” Venkat said.

MBIR is particularly useful as scientists use neutrons to evaluate objects. The method involves measuring how neutrons interact with an object’s nuclei and “bounce” away, or scatter, revealing information about the object’s inner structure. “The challenge is, how do we take those measurements, account for the interaction mechanism between the neutron and the material and then create an accurate 3D image of the object being observed,” Venkat said.

Using Venkat’s advanced algorithms, scientists can create 3D models with fewer images—which can both speed imaging and lower its cost. “These new algorithms can push what we can get out of the machines from a limited set of measurements,” he noted.

At ORNL, his research has been focused on areas such as new materials and methods to aid the U.S. manufacturing sector, and the inner workings of human cells to advance drug discovery.

Venkat’s research is assisting materials scientists at the DOE Manufacturing Demonstration Facility at ORNL who are working on methods to efficiently and affordably validate 3D printed parts for end use. “We’re using neutrons, but we also want to see what we can do to push the capabilities of x-ray machines,” Venkat said. “We can use computation to improve the quality of images without the expense of new hardware and cameras.”

Likewise, he is participating in a laboratory-directed research and development project in which ORNL is using measurements of proteins supplied by the cryo-electron microscopy instrument at Frederick National Laboratory to produce 3D images and verifying the structure using neutrons and simulations that rely on high-performance computing resources. 

“The role of computation is still growing. Scientists in many fields like manufacturing and in drug research didn’t start out thinking about computation and imaging. But enabling areas like better microscopy can allow us to make better materials and find new, more effective treatments for disease,” Venkat said. “This is an enabling tool for multiple technologies.”

Growing up, Venkat spent time in his homeland of India, as well as in the Middle East where his father was a logistics manager for the crown prince of Abu Dhabi. His father passed on a natural curiosity about how things work to Venkat.

In high school, Venkat was keen on physics, math, and computer science. He combined those interests to earn a bachelor’s degree in electrical engineering at the National Institute of Technology in India, where he also developed a fascination with signal processing—a field that models and analyzes data representations of physical states.

A team approach to advancing science

As an undergrad he worked on a speech processing project, developing a biometric system that sought to authenticate people by their voices. “This kind of signal processing involved a lot of applied mathematics and an understanding of physics to solve real-world issues, and that was very exciting to me,” he noted.

Venkat later chose Purdue University for his graduate studies because of its strong program where those same signal processing concepts were applied to imaging. He earned master’s and doctoral degrees in electrical and computer engineering from Purdue, with a thesis topic on MBIR for micro- and nanoscale tomography.

In between his master’s and PhD studies at Purdue, Venkat worked in the oil industry—on signal processing to guide drill bits exploring underground reservoirs for oilfield services firm Baker Hughes.

After earning his PhD, Venkat continued his 3D imaging work at Lawrence Berkeley National Lab. During that time, he visited ORNL as a user of the Oak Ridge Leadership Computing Facility. Venkat later learned about a Wigner Fellowship opportunity in the Imaging, Signals, and Machine Learning Group at ORNL from a member of the group who also happened to be a former Purdue alum.

Away from the lab, Venkat follows cricket and plays the game recreationally. He also enjoys reading and listening to podcasts. Lately his casual reading has centered on the link between the mind and human behavior.

Venkat touts ORNL as a great place to work. “It’s exciting to work with some of the biggest science tools available, and with a team made up of technically strong, and just very nice people. That sounds like a cliché, but it’s important. In the field I’m in, I can’t just sit and work on my own. I’m tied to application areas.”

Venkat said he enjoys working with a team for the greatest scientific impact. “That’s what I find most exciting: trying to solve a real-world, impactful problem. It’s a driving force for me. I want to help scientists get the information they need in the fastest way possible. If we have a software tool that can work with instruments to make them faster and more reliable—that’s a huge advancement.”

ORNL is managed by UT-Battelle for the Department of Energy's Office of Science, the single largest supporter of basic research in the physical sciences in the United States. DOE’s Office of Science is working to address some of the most pressing challenges of our time. For more information, please visit http://science.energy.gov.—Stephanie Seay

Share