AMM

AMM represents piecewise multilinear volumetric data using mixed-precision adaptive meshes to reduce in-memory and on-disk footprints of uniformly sampled scalar fields.

Key Features:

• Resolution-Precision Adaptivity: Simultaneously adjusts spatial resolution and numerical precision to achieve enhanced data compression compared to applying either strategy independently.
• Flexible Spatial Hierarchy: Manages spatial hierarchy by selectively relaxing or enforcing constraints on conformity, continuity, and coverage to provide an adaptive representation.
• Mixed-Precision Representation: Encodes values with mixed numeric precision to reduce storage while preserving essential details in different dataset regions.
• Incremental Creation: Supports incremental construction of AMMs from arbitrary orderings of input data to facilitate processing of large datasets.
• Interoperability with VTK: Interfaces with the Visualization Toolkit (VTK) to enable integration with existing rendering and visualization workflows.

Scientific Applications:

• Medical imaging: Compactly represents high-resolution volumetric medical scans while preserving diagnostically relevant structures.
• Geospatial analysis: Reduces storage and memory demands for large-scale 3D geospatial datasets and terrain models.
• Scientific simulations: Stores and transmits large simulation scalar fields with reduced footprint while retaining essential features.
• Visualization of high-resolution scalar fields: Enables rendering and analysis of complex scalar fields and high-resolution 3D models with reduced data size.

Methodology:

Creates a piecewise multilinear representation of uniformly sampled scalar data and uses adaptive mechanisms to dynamically adjust spatial resolution and numerical precision; supports incremental construction from arbitrary data orderings and a flexible spatial hierarchy by relaxing or enforcing conformity, continuity, and coverage constraints.