Dynamical Neural Networks: modeling low-level vision at short latencies
Synopsis: This paper is a review paper focusing on the parallel and event-based nature of neural processing in low-level vision. It shows that using correlation-based lateral interactions, one could better use regularities in natural images. This formalization opens the way to applications in image processing and more generally of using efficient probabilistic distributed representations with greedy algorithms.
Get a reprint
this chapter is included in the book Topics in Dynamical Neural Networks
- Laurent Perrinet. Dynamical Neural Networks: modeling low-level vision at short latencies, URL . pages 163--225. abstract
All material (c) L. Perrinet. Please check the copyright notice.
This work was supported by European integrated project FP6-015879, "FACETS".