Trick-GS: A Balanced Bag of Tricks for Efficient Gaussian Splatting

2025-01-24

AI Summary

To address the challenges of large model size, slow training, and inefficient rendering of Gaussian Splatting (GS) on resource-constrained edge devices (e.g., smartphones), this paper proposes the first lightweight GS framework tailored for edge deployment. Our method introduces three core innovations: (1) a progressive multi-scale training strategy that jointly controls image resolution, noise injection, and Gaussian scale; (2) a saliency-driven joint pruning mechanism for Gaussians and spherical harmonic (SH) bases, coupled with SH frequency-band masking; and (3) co-optimization of GPU memory usage and computational pipeline. Evaluated on three standard benchmarks, our framework achieves a 2× speedup in training, reduces model size by 40×, and doubles rendering throughput—while preserving reconstruction fidelity comparable to the original GS. This work marks the first practical realization of high-fidelity Gaussian Splatting on mobile platforms.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Gaussian splatting (GS) for 3D reconstruction has become quite popular due to their fast training, inference speeds and high quality reconstruction. However, GS-based reconstructions generally consist of millions of Gaussians, which makes them hard to use on computationally constrained devices such as smartphones. In this paper, we first propose a principled analysis of advances in efficient GS methods. Then, we propose Trick-GS, which is a careful combination of several strategies including (1) progressive training with resolution, noise and Gaussian scales, (2) learning to prune and mask primitives and SH bands by their significance, and (3) accelerated GS training framework. Trick-GS takes a large step towards resource-constrained GS, where faster run-time, smaller and faster-convergence of models is of paramount concern. Our results on three datasets show that Trick-GS achieves up to 2x faster training, 40x smaller disk size and 2x faster rendering speed compared to vanilla GS, while having comparable accuracy.

Problem

• Gaussian Splattering
• Resource-constrained Devices
• 3D Model Optimization

Innovation

• Trick-GS
• Efficiency Enhancement
• Resource-constrained Environments