Independent components analysis for representation interest point descriptors

Abstract

This paper presents a new interest point descriptors representation method based on independent components analysis (ICA). The aim of this algorithm is to find a meaningful image subspace and more compact descriptors. Combination the descriptors with an effective interest point detector, the proposed algorithm has a more accurate matching rate besides the robustness towards image deformations. The proposed algorithm first finds the characteristic scale and the location for the interest points using Harris-Laplacian interest point detector. We use Haar wavelet transform on the neighborhood of the interest points and get low frequency gradient feature vectors. Then ICA is used to model the subspace and reduces the dimension of the feature vectors. The experiments show the efficiency of the proposed algorithm. © Springer-Verlag Berlin Heidelberg 2006.