swMATH ID: 21416
Software Authors: Peer-Timo Bremer
Description: ADAPT: Adaptive Thresholds for Feature Extraction. Threshold-based feature definitions remain one of the most widely used and most intuitive choice in a wide range of scientific areas. However, it is well known that in many applications selecting a single optimal threshold is difficult or even impossible. Common examples of this are quantities with locally exponential behavior like the scalar dissipation rate in turbulent combustion simulations or indicator functions, such as, vorticity or delta popular in defining vortices. In these cases, local thresholds defined, for example, using contour trees can provide significantly better results, but typically require a data analysis expert to create and use. We present the ADAPT framework a new open source code that transforms a given scalar field such that global thresholds in the resulting field correspond to a variety of local thresholds in the original data. Consequently, the resulting field can be easily processed using any of the existing tool chains and provides scientists easy access to a wide variety of more advanced feature definitions. Currently, the package provides two techniques to define local thresholds: The relevance transform originally developed for combustion analysis and a model based fit initially designed to extract eddies in global ocean simulations. Furthermore, it provides an extendable API to easily add other transforms
Homepage: https://rd.springer.com/content/pdf/10.1007%2F978-3-319-44684-4_18.pdf
Related Software: KnotProt; MScomplex; MOAB; Halfedge Data Structures; Chimera; PyMOL; Blender; NumPy; SciPy; HAP; GAP; CGAL
Cited in: 1 Publication

Citations by Year