Event Details
Gabor's Uncertainty Principle in Visual Perception and Adaptation
- Event Date: December 11, 2008
- Event Start Time: 12:00 PM
- Event End Time: 7:00 PM
- Event Location: Salk Institute for Biological Studies, USA;
- Event Type: Human and Computer Vision Series
- Event Semester: Fall 2008
- Event Contact: Sergei Gepshtein
- Event Extra info: <a href="http://www.vcl.salk.edu/~gepshtein/">Sergei Gepshtein</a>
I will present a new theory of sensitivity in a visual system constrained by Gabor's uncertainty principle of measurement. According to the principle, the precision of estimating signal location and the precision of estimating signal content are not independent of one another. At the limit, precision of one kind of measurement can be improved only in expense of precision of the other kind.
It has been argued that this tradeoff of precision determines properties of receptive fields in the primary visual cortex. I will show that the uncertainty principle has much more far-reaching consequences for vision than the local computations: A visual system constrained by the uncertainty principle must have a number of invariant global properties which are strikingly similar to the known characteristics of biological vision.
The talk will consist of two parts. First, I will show how this approach explains the peculiar shape of the spatiotemporal contrast sensitivity surface, resolves inconsistencies in the literature on apparent motion, and helps to understand why motion adaptation sometimes improves and sometimes impairs visual performance. Second, I will present new results on motion adaptation supporting predictions of a normative theory of adaptation driven by the uncertainty principle.