This paper introduces a novel networking strategy for RAM- based Neurons which significantly improves the training and recognition performance of such networks while maintaining the generalization capabilities achieved in previous network configurations. A number of different architectures are introduced each using the same underlying principles. Initially, features which are common to all architectures are described illustrating the basis of the underlying paradigm. Three architectures are then introduced illustrating different techniques for employing the paradigm to meet differing performance specifications. The architectures are described in terms of the structure of the neurons they employ. Greater detail of the various training and recognition algorithms employed by the architectures may be found in the referenced papers.

Date Published: 31 October 1996
PDF: 8 pages
Proc. SPIE 2908, Machine Vision Applications, Architectures, and Systems Integration V, (31 October 1996); doi: 10.1117/12.257276

Published in SPIE Proceedings Vol. 2908:
Machine Vision Applications, Architectures, and Systems Integration V
Susan Snell Solomon; Bruce G. Batchelor; Frederick M. Waltz, Editor(s)