Spectral three dimensional mesh matching using histograms descriptors and salient points

Abstract

Recently, 2-D and 3-D shape matching have known a great interest in image and mesh processing. We propose a new spectral method based on the study of the behavior of the dual graph Laplacian matrix of the 3D mesh. Two aspects are evoked in our approach, the first one considers salient detected points matching. The second one considers the correspondEnce of the Laplacian eigenfunction histograms. Obtained results show the performances of the proposed method not only for original models but also for the corrupted ones.