TED-4DGS: Temporally Activated and Embedding-based Deformation for 4DGS Compression

2025-12-05

AI Summary

To address the low compression efficiency and poor rate-distortion (RD) performance of dynamic 3D Gaussian splatting (4DGS), this paper proposes the first RD-optimized compression framework tailored for 4DGS. Methodologically, it introduces sparse anchor points and a learnable temporal activation mechanism to enable explicit spatiotemporal modeling of Gaussian parameters; integrates lightweight temporal embeddings with a shared deformation dictionary—unifying the strengths of both static 4DGS and deformable 3DGS representations; and designs an implicit neural hyperprior model coupled with a channel-wise autoregressive entropy model to support fine-grained joint RD optimization. Evaluated on multiple real-world dynamic scene datasets, the method significantly outperforms existing 4DGS compression approaches, achieving state-of-the-art RD performance. This work establishes a new paradigm for efficient dynamic 3D scene representation.

Technology Category

• Computer Vision: 3D Computer Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• 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
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Building on the success of 3D Gaussian Splatting (3DGS) in static 3D scene representation, its extension to dynamic scenes, commonly referred to as 4DGS or dynamic 3DGS, has attracted increasing attention. However, designing more compact and efficient deformation schemes together with rate-distortion-optimized compression strategies for dynamic 3DGS representations remains an underexplored area. Prior methods either rely on space-time 4DGS with overspecified, short-lived Gaussian primitives or on canonical 3DGS with deformation that lacks explicit temporal control. To address this, we present TED-4DGS, a temporally activated and embedding-based deformation scheme for rate-distortion-optimized 4DGS compression that unifies the strengths of both families. TED-4DGS is built on a sparse anchor-based 3DGS representation. Each canonical anchor is assigned learnable temporal-activation parameters to specify its appearance and disappearance transitions over time, while a lightweight per-anchor temporal embedding queries a shared deformation bank to produce anchor-specific deformation. For rate-distortion compression, we incorporate an implicit neural representation (INR)-based hyperprior to model anchor attribute distributions, along with a channel-wise autoregressive model to capture intra-anchor correlations. With these novel elements, our scheme achieves state-of-the-art rate-distortion performance on several real-world datasets. To the best of our knowledge, this work represents one of the first attempts to pursue a rate-distortion-optimized compression framework for dynamic 3DGS representations.

Problem

• Compresses dynamic 3D Gaussian Splatting scenes efficiently
• Unifies temporal control and compact deformation for 4DGS
• Optimizes rate-distortion performance in dynamic scene compression

Innovation

• Temporally activated embedding-based deformation for 4DGS
• Sparse anchor-based 3DGS with learnable temporal activation parameters
• INR hyperprior and autoregressive model for rate-distortion compression