Recognition of three dimensional objects using a directional coupled neural network

Abstract

This report introduces a new approach on a recognition system of three dimensional, i.e., 3-D, objects. The proposed system is based on a restricted multi-layered neural network. In order to greatly reduce the total calculation cost of preprocessing for the recognition, directional coupled nodes in a network are introduced. This report also shows that the system can recognize some 3-D objects correctly in experiments.