(The Eurographics Association, 2020) Werner, Kilian; Garth, Christoph; Frey, Steffen and Huang, Jian and Sadlo, Filip
Topological simplification of merge trees requires a user specified persistence threshold. As this threshold is based on prior domain knowledge and has an unpredictable relation to output size, its use faces challenges in large-data situations like online, distributed or out-of-core scenarios. We propose two alternative parameters, a targeted percentile size reduction and a total output size limit, to increase flexibility in those scenarios.