Metric Invariants for Networks’ Classification

We suggest an approach to the shape DNA of data based on a number of metric invariants introduced by Grove and Markvorsen that encode its essential global geometry of the given structure. First experiments on real life networks and on natural images are given to demonstrate the feasibility of this approach. Even this incipient test clearly demonstrate the efficiency of the proposed invariants in the classification and understanding of stochastic textures as opposed to man-made ones.