Lossless-INR: Lossless Volumetric Implicit Neural Representations
Authors
Kaiyuan Tang (University of Notre Dame), Daniel Burke (University of Notre Dame), Chaoli Wang (University of Notre Dame)
Presentation
- Session
- Big Data, Bigger Physics
- Time
- Friday, Nov 13, 08:27 – 08:36 (US/Eastern) · session 08:00 – 09:30
- Location
- Hall America south
Links
Sign in to access the preprint PDF.
Sign in- Download Supplemental Material
Keywords
Volume visualization; implicit neural representation; lossless volumetric representation
Abstract
Implicit neural representation (INR) methods provide continuous coordinate-to-value mappings and integrate naturally with direct volume rendering, making them attractive for representing volumetric data. However, existing INR-based approaches for volumetric data are inherently lossy, and even small reconstruction errors can propagate through rendering and downstream analysis. In this work, we explore Lossless-INR, a lossless INR framework for 3D scientific volumetric data based on bit-plane decomposition. By decomposing each voxel value into binary bit-planes, we reformulate reconstruction as per-bit binary classification, so that exact recovery reduces to predicting every bit correctly. To make this optimization tractable while keeping the representation compact, we combine an octree block-partitioning strategy that adaptively subdivides complex regions with a ternary feature-grid network whose grid entries are parameterized by a ternary set of values. Experiments on diverse volumetric datasets show that this design can achieve zero bit-error rate and bit-exact reconstruction, enabling faithful rendering and downstream analysis with a compact representation. The code is available at https://github.com/TouKaienn/Lossless-INR.
For Practitioners
This paper would be of interest to practitioners working with large-scale scientific volumetric data, including simulation scientists, scientific visualization practitioners, computational scientists, medical-imaging researchers, and HPC/data-management practitioners. They could apply the proposed lossless INR framework to represent, store, transmit, and render volumetric datasets compactly while preserving bit-exact voxel values, which is important for faithful visualization, reproducible analysis, and downstream computations sensitive to reconstruction errors.