Faculty Sponsor: 
Kellman, Phil
Department: 
Psychology
Contact Name: 
Susan Carrigan
Room Number: 
Franz Hall 2349
Description of Research Project: 
In natural settings, most objects are occluded by themselves or other objects. The light such objects project to our eyes is incomplete and fragmentary. Yet our perception is not of incomplete, fragmentary objects. Our visual system overcomes the gaps in the input it receives to generate perceptions that accurately represent the objects in our world. A primary process by which we accomplish this is contour interpolation - a process by which visible contours are connected across gaps caused by occlusion. Because occlusion is ubiquitous, understanding the mechanisms responsible for interpolation is an important goal of perceptual research. A persisting and complicated problem involves determining what inputs are used to complete objects. The problem is characterized by contradictory findings regarding the influence of local geometry versus global factors such as object symmetry. Some preliminary evidence suggests that local interpolation mechanisms and more global factors may comprise separable processes. The first is a low-level process that completes contours based on local geometry. The second is a higher-order process that uses global cues, and/or past experience, to suggest the most likely shape of a partly hidden object. The result of the first process would be a crisp, precise perceptual representation of interpolated contours. The outcome of the second process would consist at least of a decision regarding the likely symmetry or familiarity of the object, but would not include a perceptual representation with precise borders. This project utilizes a dot localization method to distinguish these dual processes experimentally.
Description of Student Responsibilities: 
At a minimum: running subjects. There is potential for more responsibility, depending on what kinds of activities the student is interested in becoming involved in. Potential responsibilities: programming (in Matlab), analyzing data, presenting in lab meetings, literature review.