TinySplat: Feedforward Approach for Generating Compact 3D Scene Representation

2025-06-11

AI Summary

Existing feed-forward 3D Gaussian Splatting (3DGS) reconstruction methods achieve high speed but incur substantial storage overhead for Gaussian parameters, and mainstream compression techniques struggle to adapt due to architectural incompatibility. This paper introduces TinySplat—the first end-to-end feed-forward 3DGS compression framework—that directly synthesizes compact, high-fidelity Gaussian representations from sparse input views, eliminating iterative optimization. Its core comprises a training-agnostic triple redundancy elimination mechanism: View Projection Transformation (VPT) reduces geometric redundancy; Visibility-Aware Basis Reduction (VABR) suppresses perceptual redundancy; and standardized video coding (H.264/H.265) eliminates spatial redundancy. Experiments demonstrate that TinySplat achieves over 100× compression of Gaussian data, attaining state-of-the-art reconstruction quality using only 6% of the storage, 25% of the encoding time, and 1% of the decoding time required by prior approaches.

Technology Category

• Computer Vision: 3D Computer Vision
• Data Mining & Knowledge Management: Data Compression
• Machine Learning: Learning on the Edge & Model Compression

Application Category

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

Abstract

The recent development of feedforward 3D Gaussian Splatting (3DGS) presents a new paradigm to reconstruct 3D scenes. Using neural networks trained on large-scale multi-view datasets, it can directly infer 3DGS representations from sparse input views. Although the feedforward approach achieves high reconstruction speed, it still suffers from the substantial storage cost of 3D Gaussians. Existing 3DGS compression methods relying on scene-wise optimization are not applicable due to architectural incompatibilities. To overcome this limitation, we propose TinySplat, a complete feedforward approach for generating compact 3D scene representations. Built upon standard feedforward 3DGS methods, TinySplat integrates a training-free compression framework that systematically eliminates key sources of redundancy. Specifically, we introduce View-Projection Transformation (VPT) to reduce geometric redundancy by projecting geometric parameters into a more compact space. We further present Visibility-Aware Basis Reduction (VABR), which mitigates perceptual redundancy by aligning feature energy along dominant viewing directions via basis transformation. Lastly, spatial redundancy is addressed through an off-the-shelf video codec. Comprehensive experimental results on multiple benchmark datasets demonstrate that TinySplat achieves over 100x compression for 3D Gaussian data generated by feedforward methods. Compared to the state-of-the-art compression approach, we achieve comparable quality with only 6% of the storage size. Meanwhile, our compression framework requires only 25% of the encoding time and 1% of the decoding time.

Problem

• Reduces storage cost of 3D Gaussian representations
• Eliminates redundancy in 3D scene compression
• Improves speed and efficiency of 3DGS compression

Innovation

• View-Projection Transformation reduces geometric redundancy
• Visibility-Aware Basis Reduction mitigates perceptual redundancy
• Off-the-shelf video codec addresses spatial redundancy