A 3d shape representation and matching approach for robotic vision

In this paper we present a novel approach to 3D shape representation and matching based on the combination of the Hilbert space filling curve and Wavelet analysis. Our objective is to introduce a robust technique that capitalizes on the localization-preserving nature of the Hilbert space filling curve and the approximation capabilities of the Wavelet transform. The idea is to use a small number of 2D images to represent a 3D object. Each 2D image is in turn represented as a 1D vector resulting in a compact and efficient representation. The technique exhibits robustness in cases of partial and occluded image matching. It is translation, scale, and stretching invariant and is robust to vertical rotation.