TSGL: Teacher-Student Graph Learning for 3DGS Compression

2026-09-29

AI Summary

This study addresses the excessive storage overhead of 3D Gaussian Splatting models by proposing a post-training compression method that requires neither retraining nor access to original images. The approach leverages teacher-student graph learning to construct geometric similarity graphs and applies the Graph Fourier Transform to concentrate signal energy into low-frequency coefficients, thereby achieving a compact representation. Furthermore, a signal-dependent geometric graph learning strategy based on decoding positions and DC coefficients is designed to operate directly on pre-trained models. Experimental results demonstrate that the proposed method attains compression ratios of 27× to 33× on standard benchmarks with a PSNR degradation below 0.6 dB, significantly outperforming existing approaches.

Abstract

3D Gaussian Splatting (3DGS) is a popular representation for novel view synthesis. However, 3DGS contains millions of Gaussian primitives, each with rich attributes, resulting in large file sizes. We propose a novel 3DGS compression method based on Teacher-Student Graph Learning (TSGL) that operates directly on a trained model, without 3DGS retraining or access to training images. Specifically, for each block of Gaussian primitives, using decoded positions and DC spherical harmonic (SH) coefficients as predictors, we learn a signal-dependent geometry graph G encoding the pairwise similarities between neighbouring Gaussians via a teacher-student model. Given G, we perform Graph Fourier Transform (GFT) on the remaining attributes, so that signal energies are predominantly projected into the low-frequency coefficients for compact representation. On three standard benchmarks, the method reaches 27x to 33x compression with less than 0.6 dB of PSNR loss, improving on recent post-training compression methods in both size and rendering quality.

Problem

• 3D Gaussian Splatting
• Compression
• Novel View Synthesis
• File Size

Innovation

• 3D Gaussian Splatting Compression
• Teacher-Student Graph Learning
• Graph Fourier Transform
• Signal-dependent Geometry Graph
• Post-training Compression