Statistical computation of salient iso-values
Shivaraj Tenginakai, Raghu Machiraju · 2002
Detection of the salient iso-values in a volume dataset is often the first step towards its exploration. An error-and-trail approach is often used; new semi-automatic techniques either make assumptions about their data [4] or present multiple criteria for analysis. Determining if a dataset satisfies an algorithm's assumptions, or the criteria to be used in an analysis are both non-trivial tasks. The use of a dataset's statistical signatures, local higher order moments (LHOMs), to characterize its salient iso-values was presented in [10]. In this paper we propose a computational algorithm that uses LHOMs for expedient estimation of salient iso-values. As LHOMs are model independent statistical signatures our algorithm does not impose any assumptions on the data. Further, the algorithm has a single criterion for characterization of the salient iso-values, and the search for this criterion is easily automated. Examples from medical and computational domains are used to demonstrate the effectiveness of the proposed algorithm.