Role of Retinocortical Processing in Spatial Vision
Abstract
The inhomogeneous retinal filtering algorithms is incorporated into a more general model that includes conformal projection of the retinal filtered outputs into cortical input images, suitable for further processing, such as Gabor filtering. Our new cortical images seem to show much less loss of information relative to the retina. We no longer restore some of the dc (zero- frequency) component that is filtered out by the (Laplacian/Gaussian) retinal receptive-field model. We also provide both right- and left-hemisphere images, joined at the fovea for easy comparison with the corresponding retinal image. Study of these cortical images is yielding new insights. Peripheral objects, while remaining otherwise relatively undistorted, will be rotated either clockwise or counterclockwise as for as + or - 90 deg in cortical coordinates if they lie above or below the horizontal meridian. This is consistent with other cortical image models, but it does not bode well for the possibility of creating a stable frame by any known array-processing operation on cortical outputs. We are now beginning the third major phase of this project: modeling cortical filtering, as by Gabor functions. It is already clear that a simple, linear convolution without further refinements is not a good model for this process. Keywords: Spatial vision; Retinocortical projection; Computational model.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1989
- Accession Number
- ADA210995
Entities
People
- Donald H. Kelly
Organizations
- SRI International