Fast Dynamic Link Matching by Communicating Synapses

This work aims to provide a framework for finding continuous mappings between patterns. One important application is the correspondence problem in vision. We concentrate here on object recognition. Two patterns, identified as image and model, can have variations in size and orientation. Dynamic link matching (DLM) developed in our lab has the advantage that it can cope with translation and other variations, including deformation, but it is too slow to be the mechanism for object recognition in adults. Our new system improves the speed of DLM, by coupling synapses directly, and by thus storing and activating previously acquired mappings. Additional mechanisms are introduced for dealing with invariance in scale and orientation.