A primary goal in understanding human vision has been to characterise the details of the initial stage of spatial filtering of the retinal image and to understand the processes of feature integration and segmentation across the image. By learning about the very limits of vision (e.g. the lowest contrast that an image or its part can be just about seen) we learn about the fundamental constraints and strategies of the human signal detection system. Following this approach, I developed the noisy energy model of contrast detection (e.g. Meese & Summers, 2012), including a thorough characterisation of contrast sensitivity across the retinae (Baldwin et al, 2012). Analytic stimuli (e.g. Meese, 2010) were used to test and develop the model, addressing several gnarly issues in the field including internal noise, the level of intrinsic uncertainty, and the forms of contrast transduction and signal integration.
The main project aim is to extend the model to test it against the modelfest data set (from the NASA website). The new version will extend the model to 2D by including multiple spatial filters and investigate how filter outputs should be combined. A critical factor to the development will be determining the boundary conditions for signal integration, to be tested with appropriate experimentation. A secondary aim is to run previously successful models of the ModelFest results on several of my own challenging (and published) data sets. A third aim is to develop new stimuli that will allow further tests and comparisons between the competing models. A final aim is to test the fully specified 2D model on external noise stimuli for which there is strong theoretical motivation.
The project involves computational modelling and psychophyical experimentation. Candidates will need strong programing and quantitative skills and a basic background in visual perception and/or experimental psychology.
Some relevant references
• Modelfest details (including stimuli) can be found here http://vision.arc.nasa.gov/modelfest/
• Baldwin, Meese & Baker (2012) The attenuation surface for contrast sensitivity has the form of a witch’s hat within the central visual field. Journal of Vision, 12(11):23, 1-17.
• Meese (2010) Spatially extensive summation of contrast-energy is revealed by contrast detection of micro-pattern textures. Journal of Vision, 10(8):14, 1-21.
• Meese & Summers (2012) Theory and data for area summation of contrast with and without uncertainty: Evidence for a noisy energy model. Journal of Vision, 12(11):9, 1-28.
• Watson & Ahumada (2005) A standard model of foveal detection of spatial contrast. Journal of Vision, 5, 717-740.