A biologically inspired connectionist architecture for directing attention to salient visual field objects

In an attempt to incorporate basic visual attention abilities into existing artificial vision systems, a neural model of the bidirectional interactions within and between the brain regions believed to be involved in human visual attention has been developed. This model currently gives an artificial vision system the ability to attend to “salient”, or “pop-out” features and objects within the vision system´s field of view. After a review of the physiology of human visual attention, a network model of the aforementioned neural interactions is presented, followed by a demonstration of its performance