Hackathon-ND, a student team from computer science and engineering, was named co-winner of the IEEE Services 2020 Hackathon.
The annual competition, sponsored by the Institute of Electrical and Electronics Engineers, focuses on providing innovative solutions to societal problems related to privacy and security.
The Notre Dame team competed with teams around the world via online presentations, using the pandemic and medical data security as inspiration for its winning project “Blind U-Net for Secure COVID-19 CT Segmentation.”
“Deep neural networks and machine learning are widely used in medical applications,” said graduate student Dewen Zeng. “The key challenge in deploying these models in the real world is handling medical data securely and efficiently.”
Security regulations on medical records and the realities of segmentation in data transfer and computation make it difficult to share medical images while also maintaining patient privacy.
“Our team proposed using blind U-NET, a secure protocol that enables encryption on all data related to a patient.
“This allows a hospital, for example, to use a third-party machine learning service provider to analyze a patient’s CT scan for quicker COVID-19 diagnosis without violating privacy regulations because the third party only sees the encrypted data and no other information about the patient.”
In addition to Zeng, the Hackathon-ND team included graduate students Qing Lu and Xinrong Hu, postdoctoral research associate Weiwen Jiang, and Song Bian of Kyoto University.
Yiyu Shi, associate professor of computer science and engineering, served as the team’s mentor.
— Nina Welding, College of Engineering