Reducing interpolation artifacts by globally fairing contours

Abstract

We propose an iterative fairing method for scalar fields reducing the artifacts of bi- and trilinear interpolation. Our method reconstructs a two-dimensional scalar field from interpolation constraints optimizing the smoothness of its contours (isolines) based on variational principles. It generalizes to the trivariate case and is used to increase the quality of data sets employing a cubic B-spline representation. In contrast to filtering methods, our approach preserves the level of detail, enhancing features supported by the data while reducing interpolation artifacts.