The Promise and the Peril of Facial Recognition
During the academic year, we run a monthly live event series called Research Uncorked.
Just like the podcast, it features informal interviews with leading scholars and scientists, although in this case, our guests are drawn exclusively from Notre Dame. Originally hosted at Ironhand Winebar in South Bend, the series has been taking place online during the pandemic.
In May, we were excited to welcome Kevin Bowyer, Schubmehl-Prein Family Professor of Computer Science and Engineering. An expert in biometrics, Kevin has not only made fundamental contributions to facial recognition technology but also pursues research on its ethical implications. He talked with us about what facial recognition can do—and, just as importantly, what it shouldn’t.
Among other topics, we covered:
- his research arguing against a concept known as “criminality from face”
- algorithmic bias in face recognition technology
- positive uses of facial recognition, including applications in healthcare settings
The event was titled “The Promise and the Peril of Facial Recognition.” You can watch host Ted Fox’s entire 30-minute conversation with Kevin below.
And if you’d like to go more in-depth on Kevin’s work, he shared with us some of the research papers he and his co-authors have recently published:
- “The ‘Criminality From Face’ Illusion” in IEEE Transactions on Technology and Society
- “How Does Gender Balance In Training Data Affect Face Recognition Accuracy?,” 2020 IEEE International Joint Conference on Biometrics
- “Issues Related to Face Recognition Accuracy Varying Based on Race and Skin Tone,” in IEEE Transactions on Technology and Society
- “Remote Pulse Estimation in the Presence of Face Masks”
May 25, 2021