Euler 2D-PCA for SAR target recognition
Liu Su, Gong Zhang · 2016
Euler-Principal Component Analysis (e-PCA) has been recently proposed and successfully applied to the classification frame works. By utilizing the robust dissimilarity measure e-PCA demonstrates better performance than standard PCA while dealing with nonlinear component analysis and suppressing outliers. In this letter, we define a two-Dimensional Euler-Principal Component Analysis (e-2DPCA) framework for SAR image processing. e-2DPCA is based on 2D image matrixes rather than 1D vector which could understand two dimensional (2D) images better and get rid of high dimensional data processing. Furthermore, we applied this algorithm to SAR target recognition. Finally, experiments on MSTAR database perform the usefulness of our method in robust classification towards different situation.