This paper presents a low rate progressive 3D mesh compression scheme for simple genus-zero 3D objects. The proposed scheme is based on signal representation using redundant expansions. 3D model representation is constructed using a Matching Pursuit algorithm, with an over-complete dictionary of atoms defined on a sphere. A specific dictionary construction is proposed, that is adapted to the characteristics of 3D models. The dictionary is built on both low-frequency atoms, and anisotropic refinement to capture singularities of the signal, living on a 2-D sphere. The novel coding method is shown to favorably compare to state-of-the-art compression schemes at low bit rate, while providing a flexible and progressive representation. It therefore represents an very interesting alternative for simple 3D model representations, especially in view-dependent or scalable applications.