Using mathematics to tackle problems in medical imaging with neural networks Meet Anwesa Dey

Anwesa Dey

Over the summer Anwesa Dey researched how effective neural networks can be in solving problems in medical applications.

From the childhood basics of multiplication tables to the complex formulas needed in the medical field, math is an important component of all science. Anwesa Dey began her passion for math in grade school through the mathematical Olympiads and has since joined the University of Utah as an applied mathematics doctorate student.

It was through her university that she heard about the National Science Foundation (NSF) Mathematical Science Graduate Internship (MSGI) Program, where she saw the opportunity to take her theoretical knowledge and apply it to real world problems.

NSF MSGI provides research opportunities for mathematical sciences doctoral students, allowing them to participate in internships at national laboratories, industries and other facilities. NSF MSGI seeks to provide hands-on experience for the use of mathematics in a nonacademic setting.

Dey spent ten weeks with her mentors Johann Rudi and Getnet Betrie focusing on neural networks, a type of machine learning based on concepts from mathematics, statistics and computer science that are designed to replicate the behavior of the human brain. Neural networks have been used successfully in everything from art to economics to healthcare. Dey used neural networks to study how mathematical models can improve techniques for medical image reconstructions.

Medical image reconstructions have traditionally been computationally demanding. However, through neural networks she found that these demanding reconstructions can be fine-tuned and sped up considerably. According to Dey, deep learning techniques were especially promising. These findings could help healthcare patients receive better and faster care.

“Our study also may affect other sciences that use neural networks to solve other problems, such as in geophysical or seismic imaging,” said Dey.

Dey presented her findings at the Argonne National Laboratory during the Summer Argonne Students’ Symposium.

This internship has shaped Dey’s future research direction for her doctorate degree. She plans on using what she learned to tackle new problems in medical imaging. Dey’s experience with her mentor Rudi led to a longer-term collaboration with him as he continues to mentor her through her doctoral research.

Dey looks back on her internship proudly, having finished with new skills in several mathematical techniques. “It’s an excellent experience,” she said.

From growing up with a love of mathematics to applying it to tough real life issues, Dey’s research into the use of neural networks in medical imaging is certain to enhance the future healthcare landscape of the United States.

The NSF MSGI Program is funded by NSF and administered through the U.S. Department of Energy’s (DOE) Oak Ridge Institute for Science and Education (ORISE). ORISE is managed for DOE by ORAU.