Use algorithms to see the world differently

Cameras see the world differently than humans. Resolution, equipment, lighting, distance, and atmospheric conditions can all impact how a person interprets objects in a photo. For Sophie Voisin, a software engineer at the Department of Energy’s Oak Ridge National Laboratory, images can reveal what human eyes can’t see, giving a different perspective to understanding how the world is changing day by day.

Technically, Voisin’s work at ORNL is all about imagery – enhancing, enhancing, analyzing and exploiting high and low altitude imagery and full-motion video from drones. Every day, satellites use cameras, called sensors, to see what is happening all over the world. The volume of images captured daily is enormous. This is where Voisin’s true technical passion comes into play.

“I like to code. I like the technical side of programming algorithms to find the right place to apply the latest and greatest research to real problems,” Voisin said. “I can immediately see if the changes made to my algorithm are applied correctly.”

Maintain speed, accuracy

Voisin has been developing algorithms for seven years to process increasingly large collections of images faster. What started as a small app has grown into 12 projects; she now leads a team of 32 as lead investigators. The team aims to use software development, scientific research, and software engineering to determine which images show changes in the landscape that may be of interest to the US government. By letting computers sort through the initial mounds of images and flag some of them, analysts can then examine and interpret which images are actually of value to decision makers.

Image processing is not an easy process. The number of raw images continues to grow as sensors capture more and more snapshots of the world. Processing each image doubles or triples its size. With large file sizes making up huge datasets, high-performance computers offer the best chance of processing images at the speed needed to make meaningful decisions. Leveraging the speed of ORNL’s high-performance computers, the Voisin team also strives to offer high confidence in the accuracy of the results.

Over the past half-decade, Voisin has adapted and maintained applications as hardware changed and graphics card capabilities increased. To prevent the algorithm from breaking, Voisin and his team test new hardware before it goes live to ensure a continuous flow of information.

Algorithms make the difference

While imagery is imagery, Voisin said, it’s the file sizes that differ. Since embarking on a PhD program in computer science and image processing, she has worked on the beamline of the High Flux Isotope Reactor, a DOE Office of Science user facility, on medical imaging for mammograms and now on geospatial images. For her, the challenge of developing and tweaking algorithms for different types of images is the exciting part.

“National security work is at a different pace than other industries. I can test algorithms to fit the data and then upload it to the project,” Voisin said. But to apply machine learning algorithms to data, computer scientists need training data.

Training data helps the model learn what to look for in an image. A new algorithm may not differentiate between a van truck and a house, as both may appear as a square. Computer scientists write algorithms to detect certain features, then train the algorithm to distinguish “square shapes” and determine which ones should be flagged for further analysis.

Continue to learn from colleagues

East Tennessee has been home to Voisin and her family since she moved to the United States from France more than a decade ago. She and her husband found the opportunity at ORNL to work on fascinating projects and make an impact with applied research.

“I came to the lab for the research opportunities, but the day-to-day work is about the people you work with, and I’m grateful to work with a fantastic team,” Voisin said. “Having good relationships with my colleagues pushes me to become a good mentor because I still learn a lot from others.”

Sharon D. Cole