NEURAL STRUCTURES FOR VISUAL-MOTION TRACKING

This paper addresses visual motion tracking by a connectionist method, and aims at showing how the flexibility and the generalization power of neural networks can enhance a tracking system's adaptiveness and ef fectiveness. The simple principle of operation widens the range of app licability. A set of tracking structures that exhibit increasing level s of integration and efficiency are described. We also show how multin etwork architectures for estimate averaging may greatly increase track ing stability. The validity of the basic mechanism was assessed on a s imple domain; however, a specific difficult testbed made it possible t o verify the effectiveness of the method.