Classification using Extended Morphological Attribute Profiles based on different feature extraction techniques

Extended Morphological Attribute Profiles (EAPs) are extension of Extended Morphological Profiles (EMPs). They are based on the more general Morphological Attribute Profiles (APs) rather than the conventional Morphological Profiles (MPs). EAPs are computed on few of the first principle components (PCs) extracted from the multi-/hyper-spectral data. In this paper, we propose to compute EAPs on features derived from supervised feature extraction techniques such as discriminant analysis feature extraction (DAFE), decision boundary feature extraction (DBFE) and non-parametric weighted feature extraction (NWFE)) instead of using unsupervised principal component analysis (PCA).