arXiv Computer Vision

Transforming harmonic coefficients for 3D splat compression

arXiv Computer Vision
Aug 31

Non-Uniform Quantisation for 3DGS Compression

The paper introduces a non‑uniform quantisation scheme designed for 3D Gaussian Splatting (3DGS) models, addressing their high bitrate demands. By applying importance‑weighted quantisation and merging, the method adapts to the data distribution and removes post‑voxelisation redundancy. Experiments on benchmark datasets show state‑of‑the‑art compression performance, and the scheme is compatible with any point‑cloud representation, positioning it as a candidate for future MPEG 3DGS standardisation.

By Bert Van hauwermeiren, Patrice Rondao Alface, Adrian Munteanu
arXiv Machine Learning
Jul 1

Drop-In Perceptual Optimization for 3D Gaussian Splatting

arXiv:2603. 23297v2 Announce Type: replace-cross Abstract: Despite their output being ultimately consumed by human viewers, 3D Gaussian Splatting (3DGS) methods often rely on ad-hoc combinations of pixel-level losses, resulting in blurry renderings.

By Ezgi Ozyilkan, Zhiqi Chen, Oren Rippel, Jona Ball\'e, Kedar Tatwawadi
arXiv Computer Vision
Aug 28

KISS-GS: 3D Gaussian Splatting Compression Kept Simple

KISS-GS is a modular compression pipeline for 3D Gaussian Splatting (3DGS) scenes that separates compression from training. It first compacts a vanilla 3DGS scene by 15.7× using state‑of‑the‑art pruning, then encodes the result into the SOG‑XT image‑based format, achieving an additional 6.6× reduction. Optional encoding‑aware fine‑tuning can further cut the size by 2.2×, yielding total reductions of 85× to 319× on standard benchmarks while enabling web‑native decoding.

By Wieland Morgenstern, Friedrich Elias Branschke, Florian Fleischmann, Adrian Szatmari, Paul Schlack, Florian Barthel, Peter Eisert, Anna Hilsmann
arXiv Computer Vision
Sep 7

Compact Neural Appearance Models for Efficient Gaussian Splatting

The paper introduces a compact neural appearance model for 3D Gaussian Splatting that replaces traditional low‑order spherical harmonics (SH) with a tiny shared MLP decoding per‑primitive latent codes. It compares SH with recent spherical appearance models, integrating all into a unified CUDA rasterizer and WebGL viewer, and demonstrates that the new neural representation reduces per‑primitive appearance storage from 192 to 28 bytes, speeds optimization by 1.3×, and improves reconstruction quality. The study also analyzes how different appearance parametrizations affect geometry recovery and the handling of non‑static scene content.

By Florian Hahlbohm, Jorge Condor, Linus Franke, Martin Eisemann, Marcus Magnor
arXiv Machine Learning
Sep 1

When 3D Gaussian Splatting Recovers Real Surfaces

arXiv:2608.30054v1 Announce Type: new Abstract: When does 3D Gaussian Splatting (3DGS) recover the true scene surface rather than just overfitting view-dependent appearance? We answer this by develop...

By Songhe Wang, David Johnathan Miller
arXiv Computer Vision
Sep 3

CC-4DGS: Computational Deformation and Point-Cloud Compression for Storage-Efficient Dynamic Gaussian Splatting

CC-4DGS introduces a storage‑efficient framework for dynamic 4D Gaussian splatting by replacing large multi‑resolution hash tables with a deterministic dense hash encoding and compact neural decoders, reducing deformation storage to 1–3 MB per scene. It also compresses canonical point‑cloud attributes through conditional autoencoding, selective quantization, and residual codebooks, achieving 3–5× reduction in point‑cloud size with negligible quality loss. The combined approach maintains real‑time rendering performance while lowering total storage to 20–30 MB, matching state‑of‑the‑art reconstruction accuracy on N3DV and Technicolor Light Field datasets.

By Kyungdae Park, Chae Eun Rhee