A subspace method for feature extraction using independent components

We propose an approach to feature extraction using independent component analysis. The proposed method is combination of a subspace method and a combining strategy. Simulations are performed for the standard databases. Results reveal the effectiveness of the proposed method.