Closed planar shape classification using nonlinear alignment

Abstract

This paper addresses the problem associated with classification of signatures of four different types of aircraft prototypes. In order to classify the signatures, Nonlinear Alignment method is proposed. This procedure is designed to pair wise generate optimally aligned signatures by back tracking along the optimal alignment path. Classification results on these prototype signatures show that this method is quite robust in classifying the signals with unequal duration, compared to nearest mean classifier. Classification results were observed for different MSSNR for both classification methods. This paper also focused on reconstructing signatures based on the alignment path.