Combining Gradient Operators and Dihedral Angle for 2D and 3D Feature Extraction

In range images, features can be determined by, for example, looking for significant changes in gradient magnitude output, or computing normal changes. We present an approach to determining features in a range image by computing the dihedral angle between two surfaces; the partial derivatives required for this computation are provided via the use of the gradient operators presented in [14]. In addition we demonstrate how this family of gradient operators can be used to find features in both intensity and range images, and the accuracy of the approach is highlighted through the use of the Figure of Merit evaluation technique on real images.