Abstract
This paper addresses the issue of curvature coding in the mammalian visual system. Considering the successful account of cortical neurons as detectors of features of visual stimuli like orientation and spatial frequency among others, we discuss the role of curvature in visual processes, review some previous investigations on curvature and present a possible biologically plausible model of curvature detection. This model requires only orientation selective neurons (simple cells) and short-range connections, both known and well-documented features of the mammalian primary visual cortex. The dependence of the model on different parameters is explored and discussed.
Original language | English (US) |
---|---|
Pages (from-to) | 117-124 |
Number of pages | 8 |
Journal | Neurocomputing |
Volume | 65-66 |
Issue number | SPEC. ISS. |
DOIs | |
State | Published - Jun 2005 |
Externally published | Yes |
All Science Journal Classification (ASJC) codes
- Computer Science Applications
- Cognitive Neuroscience
- Artificial Intelligence
Keywords
- Contextual effects
- Curvature detection
- Orientation selectivity
- Short-range connections
- Visual system