The talk is related to the inverse problem of recovering the shape of a bounded domain from its generalized polarization tensors. I will first introduce the concept of using infinitely many generalized polarization tensors as shape descriptors for general bounded domains. Then, I will present recent results on identifying an algebraic planar domain using only a finite number of its polarization tensors. The density with respect to Hausdorff distance of algebraic domains among all bounded domains invites to extend via approximation the obtained reconstruction procedure beyond its natural context. Based on this, I will present a new algorithm for shape recognition with a few numerical illustrations using synthetic data.