Abstract

Robust clustering of data into linear subspaces is a common problem. Here we treat clustering into one-dimensional subspaces that cross the origin. This problem arises in blind source separation, where the subspaces correspond directly to columns of a mixing matrix. We present an algorithm that identies these subspaces using a modied k-means procedure, where line orientations and distances from a line replace the cluster centres and distance from cluster centres of conventional k-means. This method, combined with a transformation into a sparse domain and an L1-norm optimisation, constitutes a blind source separation algorithm for the under-determined case.