SensibleVision says racial bias not present in its facial recognition technology
Facial recognition technology from SensibleVision does not perform less accurately when applied to people of different skin color, unlike the systems tested in a recent study by MIT Media Lab, the company says in an announcement.
SensibleVision says its multi-patented 3DVerify system does not rely on shadow details, but does use advanced 3D processing developed from a database of diverse images, and therefore is immune to the racial bias of other commercial facial recognition systems. Color-based or greyscale analyses of 2D images tend to lack essential data for accurate identification, particularly in difficult lighting conditions, the company says. 3DVerify, in contrast, works by identifying unique face contours.
“This is a problem that is inherent to most facial recognition systems because their basic approach is based on limited databases of images and a focus on theory rather than real-world use cases where lighting conditions vary widely,” says George Brostoff, CEO of SensibleVision. “In many ways AI suffers from the fundamental biases of ‘real intelligence’ because it reflects a lack of diverse data points or the unconscious biases of their designers and ultimately creates these kinds of race and gender discrepancies. AI is only as good as you train it to be.”
The racial bias demonstrated by some facial recognition technologies recently caught the attention of the Subcommittee on Information Technology of the House Oversight and Government Reform Committee, which is developing guidance for government adoption of AI.
Article Topics
3D face recognition | accuracy | biometric-bias | biometrics | SensibleVision
Comments