On the search of Optimal Reconstruction Resolution

In this paper we present a novel algorithm to optimize the reconstruction fromnon-uniform point sets. We introduce a statistically-derived topology controller for selectingthe reconstruction resolution of a given non-uniform point set. Deriving information fromhomology-based statistics, our topology-controller ensures a stable and sound basis for theanalysis process. By analysing our topology-controller, we select an optimal reconstructionresolution which ensures both low reconstruction errors and a topological stability of the underlyingsignal. Our approach offers a valuable method for the evaluation of the reconstructionprocess without the need of visual inspection of the reconstructed datasets. By meansof qualitative results we show how our proposed topology statistics provides complementaryinformation in the enhancement of existing reconstruction pipelines in visualization.