From a science center classroom to the halls of MIT to a conference in Japan, senior computer science major Eric Brown has seized every opportunity presented to himself and more.
Brown’s most recent research looked to find an effective way to assess pain levels in non-communicative patients, particularly those with Alzheimer’s disease. He worked with Oyster River High School student Andrew Lu, a volunteer research assistant at the University of New Hampshire (UNH), utilizing Lu’s hardware knowledge and combining it with his software knowledge to craft the final product.
However, Brown’s drive for research is not a new development. Brown’s first published paper came about after taking a machine learning course taught by Wei Lu, the computer science department chair.
The paper was an analysis of various machine learning models and their effectiveness in combating distributed denial of services (DDOS) attacks.
The report, which Brown worked on with KSC alumni John Fisher, Aaron Hudon and Erick Colston, opened the door for him to present at the 38th annual Advanced Information Networking and Applications (AINA) conference in Kitakyushu, Japan, as well as a summer internship with Wei Lu to do more work.
“It’s very rewarding doing this research and then presenting it [because] it’s such a real-world application,” Brown said of his work. “You [can] actually impact the world around you, which I think is really fascinating and rewarding.”
Brown worked on two papers over the course of his summer internship, one of which landed him and Andrew Lu at MIT for the Undergraduate Research Technology Conference on Saturday, Oct. 12.
Wei Lu said typically his role in supervising research is to help find a research direction, but with Brown, he already knew he wanted to tackle a healthcare-related topic.
“I got surprised at the beginning. I think this is really creative and amazing, right? And all of these pictures and the database … the only thing I did in this at the beginning of this project is I signed the paperwork to request the data site from the other third party source,” Wei Lu said. “[Brown] is really our top computer science student.”
Brown said Lu’s expertise and weekly check-ins helped him to stay on task with his work, but also stressed the importance of learning self-reliance.
“A large part in being a researcher is having to be self-sufficient so you know the professor isn’t always going to be there for you, especially after graduation,” Brown said.
In terms of the paper itself, Brown summarized how pain could be assessed in patients through his and Andrew Lu’s research, “The machine learning model can predict if somebody’s in pain or if somebody’s not in pain by using the facial features.”
Brown and Andrew Lu’s poster explained that the pair used a graphic user interface (GUI) with a camera, a list of images to choose from and an emotion analysis function.
The system took into account a dataset of images displaying gradual levels of pain, focusing on the placement of muscles on the face and determining lowered eyebrows and tight lips as two indicators of heightened pain levels.
“When you’re in pain, there’s no happiness involved with that, but it has a mixture of having anger, sadness, maybe disgust, but sometimes it’s also neutral,” Andrew Lu said.
The pair concluded that cropping of the images used did not significantly affect emotion classification and that accuracy of emotional recognition is not a good indicator of pain level, but rather that examining action units of the face provides a better insight into pain levels.
“We can also use the motion recognition models which often use checking the face for different muscles, also known as [action] units, which help with identifying pain,” Andrew Lu said.
Additionally, Andrew Lu spoke to his experience at MIT.
“It gives me a feeling [that] when I go into these places, I understand what other people are trying to accomplish at the moment and how technology is expanding. I can touch the outside edge of what people are currently doing, offering and developing for now,” he said.
Going forward, Brown plans to publish and submit the paper to the 2025 AINA conference in Spain and use the work he’s done to pursue a master’s degree post-graduation.
“It’s been awesome [and] I’m really grateful I took that opportunity,” Brown said.
Nathan Hope can be contacted at
nhope@kscequinox.com
