Neural Models of Motion Perception
Abstract
Eight research projects supported by this grant during the reporting period have resulted in three refereed publications, one under review, one book chapter, and three conference papers. Areas of research included design and simulation of neural architectures for: (1) multichannel data fusion; (2) object recognition and image understanding; (3) development and refinement of algorithms for segmentation, boundary completion, and featural filling-in based on BCS/FCS architectures; (4) network design and simulations of an architecture for breaking of unwanted persistence (hysteresis) of visual segmentations; (5) design of a network architecture for explaining human capabilities for efficient detection of targets in clutter; (6) design and execution of human psychophysical experiments for constraining development of the BCS; (7) design and simulation of a network architecture for enhancing featural contrast and boundary localization at line-ends and corners through a novel circuit analog of VI to lateral geniculate nucleus feedback; and (8) relation of hyperacuity and illusory contour data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 22, 1994
- Accession Number
- ADA278362
Entities
People
- Ennio Mingolla
- Stephen Grossberg
Organizations
- Boston University