Convolutional neural net face recognition works in non-human-like ways



Hancock PJB, Somai RS & Mileva VR (2020) Convolutional neural net face recognition works in non-human-like ways. Royal Society Open Science, 7 (10), Art. No.: 200595.

Convolutional neural networks (CNNs) give the state-of-the-art performance in many pattern recognition problems but can be fooled by carefully crafted patterns of noise. We report that CNN face recognition systems also make surprising ‘errors'. We tested six commercial face recognition CNNs and found that they outperform typical human participants on standard face-matching tasks. However, they also declare matches that humans would not, where one image from the pair has been transformed to appear a different sex or race. This is not due to poor performance; the best CNNs perform almost perfectly on the human face-matching tasks, but also declare the most matches for faces of a different apparent race or sex. Although differing on the salience of sex and race, humans and computer systems are not working in completely different ways. They tend to find the same pairs of images difficult, suggesting some agreement about the underlying similarity space.

Convolutional Neural Nets; Automatic Face Recognition; Human Face Matching

Royal Society Open Science: Volume 7, Issue 10

FundersEPSRC Engineering and Physical Sciences Research Council
Publication date31/10/2020
Publication date online07/10/2020
Date accepted by journal15/09/2020
PublisherThe Royal Society

People (2)


Professor Peter Hancock

Professor Peter Hancock

Professor, Psychology

Dr Viktoria Mileva

Dr Viktoria Mileva

Lecturer in Psychology, Psychology

Projects (1)

FACERVM - Face Matching

Research centres/groups

Research themes