Spacetime Gaussian Feature Splatting for Real-Time Dynamic View Synthesis

TL;DR

This paper introduces Spacetime Gaussian Feature Splatting, a novel dynamic scene representation that enables high-quality, real-time, and compact view synthesis of dynamic scenes, outperforming existing methods in both quality and speed.

Contribution

The paper presents a new dynamic scene representation using enhanced 3D Gaussians with temporal features, neural feature splatting, and guided sampling, achieving state-of-the-art results.

Findings

Achieves state-of-the-art rendering quality and speed.

Supports 8K resolution rendering at 60 FPS on high-end GPU.

Maintains compact storage while modeling complex dynamic scenes.

Abstract

Novel view synthesis of dynamic scenes has been an intriguing yet challenging problem. Despite recent advancements, simultaneously achieving high-resolution photorealistic results, real-time rendering, and compact storage remains a formidable task. To address these challenges, we propose Spacetime Gaussian Feature Splatting as a novel dynamic scene representation, composed of three pivotal components. First, we formulate expressive Spacetime Gaussians by enhancing 3D Gaussians with temporal opacity and parametric motion/rotation. This enables Spacetime Gaussians to capture static, dynamic, as well as transient content within a scene. Second, we introduce splatted feature rendering, which replaces spherical harmonics with neural features. These features facilitate the modeling of view- and time-dependent appearance while maintaining small size. Third, we leverage the guidance of training error and coarse depth to sample new Gaussians in areas that are challenging to converge with existing pipelines. Experiments on several established real-world datasets demonstrate that our method achieves state-of-the-art rendering quality and speed, while retaining compact storage. At 8K resolution, our lite-version model can render at 60 FPS on an Nvidia RTX 4090 GPU. Our code is available at https://github.com/oppo-us-research/SpacetimeGaussians.