Procrustes methods for projective shape

Consider a configuration of k points or landmarks in R, represented as a k × m matrix X0. For many purposes, it is not the configuration of X0 which is of importance but its shape, that is, its equivalence class under an appropriate group of transformations. One important type of shape is projective shape. Let X = [1k, X0] be a k × p matrix, where p = m + 1, containing the landmark positions in homogeneous coordinates. In the theory of projective shape we treat