Compressed 3D Gaussian Splatting for Accelerated Novel View Synthesis

TL;DR

This paper introduces a compressed 3D Gaussian splat representation that significantly reduces memory usage and enhances rendering speed for novel view synthesis, making it suitable for low-power devices and real-time applications.

Contribution

It proposes a novel compression method using sensitivity-aware vector clustering and quantization-aware training, achieving up to 31x compression with minimal quality loss.

Findings

Achieves up to 31x compression rate on real-world scenes.

Enables rendering at up to 4x higher framerates on lightweight GPUs.

Demonstrates robustness and efficiency across multiple datasets.

Abstract

Recently, high-fidelity scene reconstruction with an optimized 3D Gaussian splat representation has been introduced for novel view synthesis from sparse image sets. Making such representations suitable for applications like network streaming and rendering on low-power devices requires significantly reduced memory consumption as well as improved rendering efficiency. We propose a compressed 3D Gaussian splat representation that utilizes sensitivity-aware vector clustering with quantization-aware training to compress directional colors and Gaussian parameters. The learned codebooks have low bitrates and achieve a compression rate of up to $31\times$ on real-world scenes with only minimal degradation of visual quality. We demonstrate that the compressed splat representation can be efficiently rendered with hardware rasterization on lightweight GPUs at up to $4\times$ higher framerates than reported via an optimized GPU compute pipeline. Extensive experiments across multiple datasets demonstrate the robustness and rendering speed of the proposed approach.