FeatureZ : A General Framework for Feature-Preserving Compression via Pointwise Bounds and Star Classification
This paper introduces FeatureZ, a lossy compression framework for structured volumetric scalar fields that can preserve a wide class of geometric and topological features and demonstrates that FeatureZ preserves diverse features during compression with minimal overhead.
Nathaniel Gorski, Xin Liang, Han-Qi Guo et al.
· 0 citations