Abstract :
High-dimensional shape analysis in state space is proposed as a new methodology for oceanic signal classification. Highdimensional moment invariant and principal component analysis are employed as two features in this framework. Experimental results show that the proposed scheme can provide some information not contained in traditional features.
