Research Supervisor Connect

Signal processing in the visual system

Summary

Research in my laboratory focuses on signal processing in the visual system. I use three methodologies - computational modelling, the analysis of existing neuronal data, and human psychophysics. There are several ongoing projects.

Supervisor

Dr Alan Freeman.

Research location

Camperdown - School of Medical Sciences

Program type

Masters/PHD

Synopsis

Computational modelling
A recently published paper from my laboratory (Nguyen and Freeman, 2019) quantitatively describes signal processing in the upstream visual system, including subcortical pathways and primary visual cortex. It shows how subcortical pathways converge on cortical cells, how this connection changes in the developing visual system, and the resulting patterns of orientation selectivity in primary visual cortex. We are now extending the model to include motion sensitivity and mechanisms underlying binocular depth perception.

Analysis of existing neuronal data
We test the model not only on published results but also on data recently collected from mammalian primary visual cortex in a colleague's laboratory (Luo-Li, Mazade, Zaidi, Alonso and Freeman, 2018).

Human psychophysics
Another way of testing the model is to compare its output with responses from human subjects viewing visual stimuli in the laboratory. We use the same stimuli computationally and psychophysically to find the extent to which modelling can account for perceptual data (Luo-Li, Alais and Freeman, 2016).

Machine vision
The computational model that we have built is potentially very general. It currently reproduces several important visual capabilities - such as orientation selectivity - and can potentially display motion direction selectivity. The ultimate goal here is to provide the front end of a machine-based object recognition system. The most powerful system resides in the human brain, and it therefore makes sense to copy the biological version as closely as possible.



Additional information

.

Want to find out more?

Opportunity ID

The opportunity ID for this research opportunity is 1167