Facial recognition might work better with a dedicated capture device, like all other biometrics
Facial recognition would probably work better with images taken from a dedicated device, rather than one designed to capture good pictures of unrelated objects like landscapes, Arun Vemury said during a recent episode of the ‘Technologically Speaking’ podcast.
The Department of Homeland Security’s Science and Technology Directorate’s (S&T’s) podcast brought in the biometrics expert for a discussion focused on facial recognition.
Vemury is the director of DHS’s Biometric and Identity Technology Center, and he talked about the role of biometrics in the “three-legged stool” of identity and issues related to facial recognition with host John Verrico.
Their conversation covered the massive gains in algorithm design and performance that have made facial recognition so much more accurate than just a few years ago, and the potential to differentiate between identical twins with face biometrics. They delved into challenges of facial recognition around perception of the technology and overgeneralization, algorithmic bias and what constitutes acceptable variations in performance.
Then the discussion turned from algorithms to the cameras that capture images used in facial recognition.
“I think part of the challenge is, we’ve had, we take cameras, and we use cameras that are kind of general purpose, right?” Vemury points out. “The cameras we buy, can take photos of landscapes, can take photos of birds, can take photos of potted plants, any number of things. I think we need cameras that are designed to take really good photos of faces, and really better accommodate the diversity of human faces and human skin tones.”
This argument draws on direct equivalents in other biometric modalities.
“If you think about it for fingerprints, you know, we don’t just use any old device for a finger,” Vemury says. “We have dedicated scanners and devices to take a fingerprint. If I were going to use my irises, I would go to a dedicated camera that is really good at taking iris photos. We need to do the same thing probably if we want to get to these really great levels of matching performance or facial recognition cameras too.”
Vemury says he sees greater attention to camera technology emerging lately in the tech and security industries.
The discussion moved on from there to the privacy implications of facial recognition and benefits of mDLs.
Google parent Alphabet released a set of tuning models and algorithms to better capture different skin tones earlier this year.