Understanding Face Processing

The human face contains a rich source of sensory information that humans use to recognise one another. While humans are highly skilled at recognising faces under differing conditions (e.g. pose, expression, lighting), many computer algorithms struggle to recognise faces when they differ from the traditional ‘passport-style’ pose. The Hummingbird Project will explore the factors contributing to good and poor facial recognition by humans and computer algorithms. This will allow us to improve performance of computer algorithms by implementing human-like strategies for facial recognition.