VIU LogoVision & Image Understanding

Face Perception

face FIO

Each color map shows predicted accuracy (red high performance, blue low) for two different model predicting the 1st eye movement of humans to faces. Left: A region of interest ideal observer which predicts eye movements to the most informative featues in the face (the eyes);. Right: A foveated ideal observer model constrained by the degrading visual processing with retinal eccenticity. This model predicts that the best place to initially look at a face is just below the eyes. Green circles are average 1st saccades for twenty five observers. White circle average location across observers.

Looking at a face to recognize a person's identity, their emotional state, their intention is arguably among the most practiced and important perceptual tasks for humans. The human brain can extract all of this inforamtion very briefly. Our research aims at uncovering: the sophisticated brain algorithms that allow for this incredibly fast and automatic face recogntion, the differences across individuals on how they look at faces, and their neural correlates.


Or, C. C. F., Peterson, M. F., & Eckstein, M. P. (2015). Initial eye movements during face identification are optimal and similar across cultures.Journal of Vision15(13), 12-12

Peterson, Matthew F., and Miguel P. Eckstein. "Learning optimal eye movements to unusual faces." Vision research 99 (2014): 57-68.

Peterson, M. F., & Eckstein, M. P. (2013). Individual Differences in Eye Movements During Face Identification Reflect Observer-Specific Optimal Points of Fixation. Psychological science

Peterson, M. F., & Eckstein, M. P. (2012). Looking just below the eyes is optimal across face recognition tasks. Proceedings of the National Academy of Sciences, 109(48)

Das, K., Giesbrecht, B., & Eckstein, M. P. (2010). Predicting variations of perceptual performance across individuals from neural activity using pattern classifiers. Neuroimage, 51(4), 1425-1437

Peterson, M. F., Abbey, C. K., & Eckstein, M. P. (2009). The surprisingly high human efficiency at learning to recognize faces. Vision research, 49(3), 301-314.