Abstract:3D Gaussian splatting (3DGS) has recently emerged as an alternative representation that leverages a 3D Gaussian-based representation and introduces an approximated volumetric rendering, achieving very fast rendering speed and promising image quality. Furthermore, subsequent studies have successfully extended 3DGS to dynamic 3D scenes, demonstrating its wide range of applications. However, a significant drawback arises as 3DGS and its following methods entail a substantial number of Gaussians to maintain the high fidelity of the rendered images, which requires a large amount of memory and storage. To address this critical issue, we place a specific emphasis on two key objectives: reducing the number of Gaussian points without sacrificing performance and compressing the Gaussian attributes, such as view-dependent color and covariance. To this end, we propose a learnable mask strategy that significantly reduces the number of Gaussians while preserving high performance. In addition, we propose a compact but effective representation of view-dependent color by employing a grid-based neural field rather than relying on spherical harmonics. Finally, we learn codebooks to compactly represent the geometric and temporal attributes by residual vector quantization. With model compression techniques such as quantization and entropy coding, we consistently show over 25x reduced storage and enhanced rendering speed compared to 3DGS for static scenes, while maintaining the quality of the scene representation. For dynamic scenes, our approach achieves more than 12x storage efficiency and retains a high-quality reconstruction compared to the existing state-of-the-art methods. Our work provides a comprehensive framework for 3D scene representation, achieving high performance, fast training, compactness, and real-time rendering. Our project page is available at <a class="link-external link-https" href="https://maincold2.github.io/c3dgs/" rel="external noopener nofollow">this https URL</a>.

GSCore: Efficient Radiance Field Rendering via Architectural Support for 3D Gaussian Splatting

Efficient Density Control for 3D Gaussian Splatting

Compact 3D Gaussian Splatting for Static and Dynamic Radiance Fields

3D Convex Splatting: Radiance Field Rendering with 3D Smooth Convexes

Identifying Unnecessary 3D Gaussians using Clustering for Fast Rendering of 3D Gaussian Splatting

AdR-Gaussian: Accelerating Gaussian Splatting with Adaptive Radius

GES: Generalized Exponential Splatting for Efficient Radiance Field Rendering

RadSplat: Radiance Field-Informed Gaussian Splatting for Robust Real-Time Rendering with 900+ FPS

3D Gaussian Splatting for Real-Time Radiance Field Rendering

FlashGS: Efficient 3D Gaussian Splatting for Large-scale and High-resolution Rendering

Compact 3D Gaussian Representation for Radiance Field

Isotropic Gaussian Splatting for Real-Time Radiance Field Rendering

GS-Octree: Octree-based 3D Gaussian Splatting for Robust Object-level 3D Reconstruction Under Strong Lighting

2D Gaussian Splatting for Geometrically Accurate Radiance Fields

GS‐Octree: Octree‐based 3D Gaussian Splatting for Robust Object‐level 3D Reconstruction Under Strong Lighting

Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives

AtomGS: Atomizing Gaussian Splatting for High-Fidelity Radiance Field

EAGLES: Efficient Accelerated 3D Gaussians with Lightweight EncodingS

GS^3: Efficient Relighting with Triple Gaussian Splatting

GS3: Efficient Relighting with Triple Gaussian Splatting

Taming 3DGS: High-Quality Radiance Fields with Limited Resources