A NEW NEURAL MODEL FOR INVARIANT PATTERN-RECOGNITION

For most of the pattern recognition applications, it is often required to correctly recognize patterns even if they have variations in posit ion, rotation, and/or scale. In this paper, to achieve the goal of inv ariant pattern recognition we propose a new neural model which consist s of a cascade connection of four two-dimensional layers. The first th ree layers of the neural model perform the processes of position norma lization, rotation normalization and feature extraction, respectively. The last layer is responsible for both recognition job and scale norm alization by specially designing its output neurons to possess a scale invariant property. Finally, simulation results are given to demonstr ate that the proposed model is simple and effective for invariant patt ern recognition. Copyright (C) 1996 Elsevier Science Ltd