July 21, 2017 – At the heart of Hector Santos-Villalobos’ work is the need to get a better look at the unknown, whether it is the structure of an underground geothermal reservoir, meter-thick concrete surrounding a nuclear reactor, or the fine details of the human eye. It may seem appropriate, then, to bring his research and the path that led him to it into focus.
Santos-Villalobos' work in computational imaging and biometrics at the Department of Energy's Oak Ridge National Laboratory involves using conventional technology like x-ray machines and high-resolution camera lenses and applying novel algorithms and data analysis to sharpen resulting images.
“Most people don’t have a $1 million instrument to show them exactly what they want. So we figure out how to modify or enhance images from conventional systems to accomplish what they need,” said Santos-Villalobos, a scientist in the Imaging, Signals, and Machine Learning group in the Electrical and Electronics Systems Research Division.
Santos-Villalobos' non-destructive evaluation work at the lab involves the development of techniques such as ultrasonic linear arrays to help researchers peer into the thick concrete containment walls used in nuclear power plants without disturbing them. “It’s hard to penetrate these systems. Traditional imaging can result in many reflections and ‘noise’ that create a distorted view,” he said.
Researchers needed to come up with a way to overcome those distortions, and they needed a way to do so quickly. Nuclear power plant operators, for instance, cannot wait for months for images after measurements are taken during inspections, Santos-Villalobos said. What he and collaborating researchers at ORNL and elsewhere developed is a model for the images that takes into account the physics of the instrument and the materials known to be present in a specimen—reinforcing rebar for instance—to more clearly identify other unknown objects such as potential defects. The technique is called a model-based iterative reconstruction (MBIR) algorithm.
Similarly, Santos-Villalobos and his colleagues have studied the use of ultrasound and MBIR to evaluate the integrity of wellbores and associated underground reservoirs, whether geothermal energy or oil and gas deposits. The work can be used to ensure wells do not leak and to make extraction of such energy resources more efficient.
In his biometrics work, Santos-Villalobos is researching the use of plenoptic computational photography to improve the recognition of human irises. He and his colleagues have developed plenoptic camera arrays that capture light fields to give a better 3D image of the object being photographed.
Backyard experiments
Santos-Villalobos said he was always interested in science and the “why” of things as a child growing up in Puerto Rico. When he was about 10, Santos-Villalobos recalls trying to test a hypothesis that lizards were descendants of dinosaurs. He tried to create a chemical reaction that would release their dinosaur DNA—using his mother’s laundry detergent. The experiment resulted in an annoyed but otherwise unharmed backyard lizard, but it did not stifle young Santos-Villalobos' scientific curiosity.
Santos-Villalobos said he was encouraged in these endeavors by an influential middle-grades science teacher in his hometown. Even after that adviser unexpectedly passed, Santos-Villalobos took his encouragement to explore science as a career to heart.
Santos-Villalobos also credits his parents—a father who worked as a pharmaceutical technician and a mother who was a social worker—for supporting his natural curiosity and critical thinking. Santos-Villalobos was the first to earn a doctoral degree in his family.
That support was crucial for a young man who, as Santos-Villalobos describes it, grew up in a very small town where academic achievement was not as valued as prowess in other areas. “I hid my grades back then,” Santos-Villalobos said. “It was just something you didn’t talk about. So while the desire was there, I probably didn’t pursue some science questions that I may have otherwise growing up.”
It wasn’t certain that Santos-Villalobos would pursue studies at a university. During Santos-Villalobos' senior year of high school, his parents quit their jobs to start a bakery—a business that became very popular in the town and kept Santos-Villalobos and his family busy.
As a college freshman, Santos-Villalobos briefly considered leaving school to help his family grow their bakery business. But his father urged him to continue, and told Santos-Villalobos that if he wanted to work at the bakery, “I would have to work harder than anyone else.” Santos-Villalobos said. Continuing college, Santos-Villalobos adds, “was the best decision ever.”
Santos-Villalobos took that strong work ethic into the academic world. He earned both a BS and an MS in computer engineering at the University of Puerto Rico. He received a doctoral fellowship to continue his studies at Purdue University, where he earned a PhD in electrical and computer engineering. During his college career, he worked two internships with Hewlett-Packard on imaging projects.
While at Purdue, Santos-Villalobos was a research assistant and enjoyed helping fellow students as president of the Puerto Rican Student Association.
Family at home and at work
Santos-Villalobos came to ORNL in 2010 as a postdoctoral fellow, focused on processing of retinal and iris images and the design and development of a high-resolution coded source neutron imaging system. Now as an R&D staff associate, Santos-Villalobos has continued that research and has taken up mentoring of postdocs and graduate students. Last year he helped supervise several summer students.
Away from the lab, Santos-Villalobos' focus is on family and faith. He is father to a 7-month-old boy, and he and his wife work closely with their local church, where Santos-Villalobos volunteers as a translator supporting mission trips to Central America. He looks forward to the time when his son is old enough to be introduced to hiking in East Tennessee.
Santos-Villalobos said he’d like to continue in the next several years to expand his division’s computational imaging expertise. “We try to find new ways to manipulate a modality—whether it’s microscopy, medical imaging, or photography—to extend resolution and contrast to get a sharper image in less time.
“There are many applications. We have worked on medical research, and I have done some work with people at the MDF,” DOE’s Manufacturing Demonstration Facility at ORNL. In advanced manufacturing, “the holy grail is to tell whether an additively manufactured part is of such good quality that you can use it” in real-world applications, Santos-Villalobos said. “Computational imaging is one of the tools that can get us there.”
Santos-Villalobos said he enjoys the team approach to the lab’s imaging work. “There’s always a group of people behind this work, and I like to emphasize that. Working together as a team, having that cohesion, respecting all the personalities involved, and recognizing our weaknesses and strengths—that’s when we are most successful. I like to see the people I work with as family.”
UT-Battelle manages ORNL for the DOE’s Office of Science. The Office of Science is the single largest supporter of basic research in the physical sciences in the United States, and is working to address some of the most pressing challenges of our time. For more information, please visit http://energy.gov/science/.—by Stephanie Seay
