Image and Video Analysis

On the Implementation of Visual Attention Architectures

Tales of the Disappearing Computer, June 2003.

The majority of visual attention models is based on the concept of saliency map, a two-dimensional map that encodes the saliency of objects in the surrounding world. We attempt a short review of current implementations and present our first thoughts on extending the classical saliency-based model by using motion and prior knowledge.

[ Bibtex ] [ PDF ]