Spectrally Pruned Gaussian Fields with Neural Compensation

TL;DR

This paper introduces SUNDAE, a memory-efficient Gaussian field for 3D rendering that uses spectral pruning and neural compensation to reduce memory usage while maintaining high rendering quality.

Contribution

The paper proposes a novel spectral pruning method with neural compensation for Gaussian fields, significantly reducing memory consumption without sacrificing rendering quality.

Findings

SUNDAE achieves 26.80 PSNR at 145 FPS using 104 MB memory.

Compared to vanilla Gaussian splatting, SUNDAE reduces memory from 523 MB to 104 MB.

SUNDAE maintains high rendering quality with minimal performance loss.

Abstract

Recently, 3D Gaussian Splatting, as a novel 3D representation, has garnered attention for its fast rendering speed and high rendering quality. However, this comes with high memory consumption, e.g., a well-trained Gaussian field may utilize three million Gaussian primitives and over 700 MB of memory. We credit this high memory footprint to the lack of consideration for the relationship between primitives. In this paper, we propose a memory-efficient Gaussian field named SUNDAE with spectral pruning and neural compensation. On one hand, we construct a graph on the set of Gaussian primitives to model their relationship and design a spectral down-sampling module to prune out primitives while preserving desired signals. On the other hand, to compensate for the quality loss of pruning Gaussians, we exploit a lightweight neural network head to mix splatted features, which effectively compensates for quality losses while capturing the relationship between primitives in its weights. We demonstrate the performance of SUNDAE with extensive results. For example, SUNDAE can achieve 26.80 PSNR at 145 FPS using 104 MB memory while the vanilla Gaussian splatting algorithm achieves 25.60 PSNR at 160 FPS using 523 MB memory, on the Mip-NeRF360 dataset. Codes are publicly available at https://runyiyang.github.io/projects/SUNDAE/.