MEGA: Memory-Efficient 4D Gaussian Splatting for Dynamic Scenes
Authors:
- 19 October 2025
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 27828-27838
Abstract
4D Gaussian Splatting (4DGS) has recently emerged as a promising technique for capturing complex dynamic 3D scenes with high fidelity. It utilizes a 4D Gaussian representation and a GPU-friendly rasterizer, enabling rapid rendering speeds. Despite its advantages, 4DGS faces significant challenges, notably the requirement of millions of 4D Gaussians, each with extensive associated attributes, leading to substantial memory and storage cost. This paper introduces a memory-efficient framework for 4DGS. We streamline the color attribute by decomposing it into a per-Gaussian direct color component with only 3 parameters and a shared lightweight alternating current color predictor. This approach eliminates the need for spherical harmonics coefficients, which typically involve up to 144 parameters in classic 4DGS, thereby creating a memoryefficient 4D Gaussian representation. Furthermore, we introduce an entropy-constrained Gaussian deformation technique that uses a deformation field to expand the action range of each Gaussian and integrates an opacity-based entropy loss to limit the number of Gaussians, thus forcing our model to use as few Gaussians as possible to fit a dynamic scene well. With simple half-precision storage and zip compression, our framework achieves a storage reduction by approximately $190 \times$ and $125 \times$ on the Technicolor and Neural 3D Video datasets, respectively, compared to the original 4DGS. Meanwhile, it maintains comparable rendering speeds and scene representation quality, setting a new standard in the field. Code is available at https://github.com/Xinjie-Q/MEGA.
Keywords
All Related Versions
This publication has 44 references indexed in Scilit:
- CompGS: Smaller and Faster Gaussian Splatting with Vector Quantization2024
- Per-Gaussian Embedding-Based Deformation for Deformable 3D Gaussian Splatting2024
- DynMF: Neural Motion Factorization for Real-Time Dynamic View Synthesis with 3D Gaussian Splatting2024
- Learning 3D Geometry and Feature Consistent Gaussian Splatting for Object Removal2024
- A Compact Dynamic 3D Gaussian Representation for Real-Time Dynamic View Synthesis2024
- MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-view Images2024
- GaussianImage: 1000 FPS Image Representation and Compression by 2D Gaussian Splatting2024
- FlashSplat: 2D to 3D Gaussian Splatting Segmentation Solved Optimally2024
- HAC: Hash-Grid Assisted Context for 3D Gaussian Splatting Compression2024
- TensoRF: Tensorial Radiance Fields2022