During the last ten years Professor Badcock’s research group has been investigating the mid-level processes in vision that form global object and motion descriptions by grouping local image information. The goal has been to elaborate the underlying properties of these mechanisms and to determine the manner in which those involved in motion processing interact with those contributing to form processing. This work contributes to both the broader fields of motion processing and shape coding and has been directly applied to study vision in Migraine, Autism, Glaucoma and Schizophrenia.

In many everyday situations the ability to determine the shape of an object is an essential precursor to sensible action. This applies equally to performance using human vision and many artificial object recognition systems. The human visual system is exquisitely sensitive and flexible when coding shape, accommodating a great variety of lighting situations, partially occluded contours and novel examples with minimal disruption to performance. Artificial shape coding systems are less effective and for this reason we still see human operators scanning the images of airport security scanners, for example. Over the last few years Professor Badcock’s laboratory has focused on the human ability to code shape, partly because an understanding of this fundamental human skill is of interest in its own right, but also because a detailed understanding of how the human visual system achieves its excellent performance is directly applicable to artificial shape coding and recognition systems, whether they are searching for faces in the crowd, matching faces with passport photographs at ports of entry,  identifying military hardware in a conflict zone, ensuring a camera only takes a shot when the people are smiling or identifying moving pedestrians to facilitate automatic collision avoidance in new Mercedes car models. Substantial progress has been made in that time but there are still important unresolved issues and this project is focused on extending that work to bring it to fruition.

The work employs predominately psychophysical (behavioural) methods to assess human visual behaviour, but have used electrophysiological techniques where necessary,  and has outlined many properties of the processes underlying shape coding from both static and moving targets. This has also been directly applied to help understand differences in visual processing in individuals who have regular migraine headaches, or who exhibit symptoms of Autism. In both of these cases the ability to process the global properties of objects and moving fields of objects is disrupted. This project provided the basic research necessary to understand those changes.

Collaborator/s

  • Associate Professor Mark Edwards, ANU
  • Dr Shin’ya Nishida Human and Information Science Lab, NTT Communication Science Labs, NTT Corporation