Shape Registration with Directional Data

TL;DR

This paper introduces new cost functions for shape registration that incorporate directional data like normals, improving alignment accuracy for 2D and 3D shapes over existing methods.

Contribution

It presents novel cost functions utilizing directional information for shape registration, applicable to both rigid and non-rigid transformations.

Findings

Using normal vectors improves registration accuracy.

Combining position and directional data outperforms state-of-the-art methods.

Applicable to 2D contours and 3D surfaces.

Abstract

We propose several cost functions for registration of shapes encoded with Euclidean and/or non-Euclidean information (unit vectors). Our framework is assessed for estimation of both rigid and non-rigid transformations between the target and model shapes corresponding to 2D contours and 3D surfaces. The experimental results obtained confirm that using the combination of a point's position and unit normal vector in a cost function can enhance the registration results compared to state of the art methods.