Sparse 4D anchors improve volumetric video compression with Gaussian splatting
Gaussian Splatting-based Volumetric Video Compression with Sparse 4D Anchors
Computer Vision and Pattern Recognition
Summary
Capturing realistic 3D videos that change over time usually creates huge files that are hard to store or send. The authors found a way to shrink these files by using special points called sparse 4D anchors to represent the video more simply and efficiently. Their method also shares information between these anchors, making the videos take up less space while still looking good. Tests show this new method compresses better than previous techniques on different video datasets.
What this means in practice
- •For video compression engineers: Compress immersive 3D dynamic videos more efficiently using sparse 4D anchors and Gaussian splatting to reduce storage and transmission costs.
- •For extended reality developers: Generate photorealistic dynamic 3D scenes with smaller data sizes for smoother VR/AR experiences.
Authors
Ge Gao, Siyue Teng, Chanqgi Wang, Fan Zhang, Nantheera Anantrasirichai, Jui Chiu Chiang, Wen-Hsiao Peng, David Bull
Abstract
Immersive video communication requires photorealistic, render-efficient, and compact dynamic scene representations. 3D Gaussian Splatting (3DGS) offers a promising representation, but dynamic 3DGS remains difficult to compress due to dense primitives and spatiotemporal redundancy. Anchor-based formulations improve compactness with sparse scaffolds that share geometry and appearance across primitives. However, existing designs often rely on deforming a single canonical scaffold and condition each primitive on its associated anchor in isolation, limiting their ability to handle non-local dynamics and disocclusion while under-exploiting inter-anchor correlations, particularly in motion- or texture-dense regions. To address these limitations, we propose SAGA, a volumetric video codec built upon Sparse Anchor-assisted GAussian splatting representations. SAGA represents dynamic 3D scenes using hierarchically organized sparse 4D anchors, where coordinate-based INR decoders generate fine anchors and Gaussian primitives from inter-anchor interpolations, enabling compact parameter sharing across spatiotemporal structures. For long-range dependencies among unstructured anchors, we further introduce fixed-size memory slots with orthogonality-informed updates for accurate entropy-context modeling. Experiments show that SAGA achieves strong rate-distortion performance against GIFStream, with PSNR BD-rate reductions of 80.39% and 83.94% on Neu3D and MPEG MIV, respectively.