Psychophysical and computational techniques are combined with data from visual neuroscience to study human visual information processing. The main focus of the work is on the computation of image motion by the visual system and its utilisation in grouping image regions to surfaces and perceiving the motion of objects, in particular, the human face. Another focus is temporal perception incorporating the perception of temporal synchrony and the study of the sensory processes involved in visual duration perception. The models developed are exploited as new technologies for image motion analysis and facial animation.
Over the last twenty years I have developed a computational model of image motion perception based on calculating spatio-temporal derivatives of a local Taylor expansion of the dynamic image structure. The differential operators used in the computation provide a good model of cortical simple cells. Later stages of the model can also be mapped onto the components of the motion pathway and the model explains how we can see complex motion sequences, for example, contrast and texture-defined motion. The current focus is on how the visual system combines local measures of image motion to deliver a perception of object or group motion.
My lab published one of the earliest studies of dynamic face perception, and we went on to exploit motion tracking and 3D graphics to separate facial motion from form. Currently we are developing a 3D dynamic system for facial motion capture to study facial imitation. This work has also led to a system for photorealistic performance driven facial animation.
Recently we have shown that adaptation to high frequency flicker can reduced the perceived duration of sub-second intervals containing intermediate rates of flicker in a localised region of visual space. This implicates early sensory processing in what previously was considered a high level, generic, timing system. By careful selection of stimuli to target the magnocellular visual pathway we have generated evidence for magnocellular involvement in this effect.