A neural network based position invariant pattern recognition system with a constrained hypermedia style user-interface

Introduces a method capable of recognising an aircraft from any viewpoint in uncluttered noisy scenes. The authors also describe a proposed exploitation of a recently developed hypermedia style user-interface for presenting and manipulating information, and for enabling users to specify image processing and pattern recognition operations. The paper is structured in two main parts. The first part describes the pattern recognition process and includes a description of the image processing procedures, a detailed description of the Quadtree and normalised Quadtree data structures, the theory and properties of the Octree data structure, and finally the actual method used in order to train and test the artificial neural network. The second part describes an initial presentation of the user-interface graphical environment and its corresponding sub-parts for each individual stage of the pattern recognition process