SUCCESS-GS: Survey of Compactness and Compression for Efficient Static and Dynamic Gaussian Splatting

TL;DR

The paper surveys Efficient Gaussian Splatting (3DGS/4DGS) with two main directions: Parameter Compression and Restructuring Compression, detailing static and dynamic techniques to reduce memory and compute while preserving rendering quality. It synthesizes methods ranging from pruning, quantization, and entropy coding to anchor-based, canonical deformable, and LoD hierarchical architectures, supported by datasets and evaluation metrics for fair benchmarking. The work highlights current limitations—hardware constraints, long-sequence dynamics, generalization gaps, and need for semantically-aware, user-controllable trade-offs—and outlines promising future directions toward scalable, real-time Gaussian Splatting for static and dynamic scenes.

Abstract

3D Gaussian Splatting (3DGS) has emerged as a powerful explicit representation enabling real-time, high-fidelity 3D reconstruction and novel view synthesis. However, its practical use is hindered by the massive memory and computational demands required to store and render millions of Gaussians. These challenges become even more severe in 4D dynamic scenes. To address these issues, the field of Efficient Gaussian Splatting has rapidly evolved, proposing methods that reduce redundancy while preserving reconstruction quality. This survey provides the first unified overview of efficient 3D and 4D Gaussian Splatting techniques. For both 3D and 4D settings, we systematically categorize existing methods into two major directions, Parameter Compression and Restructuring Compression, and comprehensively summarize the core ideas and methodological trends within each category. We further cover widely used datasets, evaluation metrics, and representative benchmark comparisons. Finally, we discuss current limitations and outline promising research directions toward scalable, compact, and real-time Gaussian Splatting for both static and dynamic 3D scene representation.

Figures (8)

• Figure 1: Overview of the static 3D Gaussian Splatting (3DGS) pipeline. A scene is represented as a set of 3D Gaussians with attributes including position $\boldsymbol{\mu}$, scale $\boldsymbol{s}$, rotation $\boldsymbol{r}$, spherical harmonics (SH) coefficients $\boldsymbol{h}$, and opacity $\sigma$ (Sec. \ref{['Sec. 2.1.1']}). Each Gaussian is projected into the camera coordinate system and rendered onto the image plane through a differentiable tile rasterizer (Sec. \ref{['Sec. 2.1.3']}, Sec. \ref{['Sec. 2.1.4']}). The rendering is optimized by minimizing the difference between predicted and GT images (Sec. \ref{['Sec. 2.1.5']}). To improve reconstruction quality, adaptive density control clones or splits Gaussians based on view-space gradient signals while pruning nearly transparent Gaussians (Sec. \ref{['Sec. 2.1.6']}). In addition, efficiency-oriented strategies such as pruning, quantization, and entropy coding are applied to compress Gaussian attributes (Sec. \ref{['Sec. 3.1']}).
• Figure 2: Overview of representative approaches for dynamic 3D/4D Gaussian Splatting. (a) Deformable 3DGSb8 initializes a canonical set of 3D Gaussians and models temporal motion using a time-conditioned MLP that predicts offsets in position, rotation, and scale. (b) Wu et al. ’s 4DGSb2 adopts a structured encoder–decoder design, where a spatial-temporal structure encoder (HexPlane) organizes the 4D domain $(x,y,z,t)$ into multi-plane features, which are then decoded into Gaussian attribute offsets. (c) Yang et al. ’s 4DGSb43 directly extends 3D Gaussians into the spatio-temporal domain by defining each Gaussian as a 4D primitive with a 4D mean, covariance, and appearance coefficients. During rendering, 4D Gaussians are converted into 3D Gaussians at target timestamps through temporal slicing.
• Figure 3: Overview of Parameter Compression strategies for Static 3DGS. These approaches reduce redundancy in the 3DGS representations without modifying the 3DGS b1 model architecture. (a) Pruning (Sec. \ref{['subsubsec:static pruning']}) removes redundant 3D Gaussians. (b) Attribute Pruning (Sec. \ref{['subsubsec:static attribute pruning']}) compresses specific Gaussian attributes. (c) Quantization (Sec. \ref{['subsubsec:static quantization']}) reduces Gaussian attribute bit precision. (d) Entropy Coding (Sec. \ref{['subsubsec:static entropy coding']}) compresses quantized Gaussian attributes by exploiting statistical redundancy. (e) Structured Compression (Sec. \ref{['subsubsec:static structured compression']}) organizes 3D Gaussians by spatial relationships to improve compression efficiency.
• Figure 4: Overview of Restructuring Compression strategies for Static 3DGS. These approaches fundamentally modify the 3DGS model architecture to achieve an efficient scene representation. (a) Anchor-based Hierarchical Structure methods (Sec. \ref{['subsubsec:static anchor-based hierarchical structure methods']}) introduce hierarchical representation to 3DGS b1 using sparse anchors. (b) Neural Network Integration methods (Sec. \ref{['subsubsec:static neural network integration methods']}) replace 3DGS representations with neural networks. (c) Geometric Structure-aware methods (Sec. \ref{['subsubsec:static geometric structure-aware methods']}) exploit geometric properties.
• Figure 5: Overview of Parameter Compression strategies for Dynamic 3DGS. These approaches reduce redundancy in the Gaussian representation without modifying the rendering architecture. (a) Attribute Pruning (Sec. \ref{['subsubsec:attribute pruning']}) removes temporally inactive components from 4D Gaussians, reducing them to 3D Gaussians that capture the invariant spatial attributes. The pruned 3D Gaussians are then combined with time-dependent 4D Gaussians at specific timestamps to reconstruct dynamic scenes while preserving spatial fidelity. (b) Gaussian Pruning (Sec. \ref{['subsubsec:gaussian pruning']}) discards less-contributing Gaussians based on temporal importance, followed by Quantization (Sec. \ref{['subsubsec:quantization']}), which discretizes Gaussian parameters to achieve compactness while preserving rendering quality.