FLUIDSPLAT uses K anisotropic Gaussian primitives as an interpretable scaffold for flow-field reconstruction, proves an O(K^{-s/d}) approximation rate under Sobolev smoothness s, derives an optimal K scaling with observation count N and noise level, and reports 11-23% error reduction on AirfRANS.
Advances in Neural Information Processing Systems , volume=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
FLUIDSPLAT: Reconstructing Physical Fields from Sparse Sensors via Gaussian Primitives
FLUIDSPLAT uses K anisotropic Gaussian primitives as an interpretable scaffold for flow-field reconstruction, proves an O(K^{-s/d}) approximation rate under Sobolev smoothness s, derives an optimal K scaling with observation count N and noise level, and reports 11-23% error reduction on AirfRANS.