Estimating non-rigid motion from point and line correspondences

Abstract

This paper presents a new algorithm for recovering motion parameters of non-rigid objects using both point and line correspondences. It requires estimating the coefficients of the first fundamental form before and after the motion. This algorithm has several advantages over a previously developed algorithm which uses only point correspondences and Gaussian curvature. First, it does not require any assumption on the spatial distribution of the stretching parameter in conformal motion. Second, the amount of computation is significantly reduced. The algorithm is tested on both simulated and real data and its performance is evaluated. Noise sensitivity of the algorithm is also evaluated using Gaussian noise on simulated data.