Feature based shape recognition using Hopfield neural network

Abstract

A key problem for robots is to identify the industrial parts in its workcell. Presently, robot workcells have limited flexibility because they expect objects in precise location without any part overlapping or touching. A method to recognize two dimensional objects independent of their position, orientation, size and limited occlusion using a Hopfield neural network is implemented. Features used are angle of variation and sphericity. The system is capable of identifying single at well as multiple occluded objects.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>