EvoGS: Constructing Continuous-Layered Gaussian Splatting with Evolution Tree for Scalable 3D Streaming

arXiv:2606.07179v1 Announce Type: new Abstract: Streaming 3D Gaussian Splatting requires highly scalable, progressive representations. Existing progressive methods rely on extit{discrete layering}, accumulating separate splat sets for each level of detail. This structural independence between layers inherently leads to error accumulation, severe splat redundancy, and uncontrolled quality transitions. We propose EvoGS, the first extit{continuous-layering} representation. Organized as an Evolution Tree, EvoGS generates finer details via an explicit, wavelet-inspired parent-child refinement. This empowers child nodes to structurally correct ancestral errors, yield inherently sparse and highly compressible inter-layer signals. Extensive experiments show EvoGS eliminates splat redundancy from over 65% to under 25%. Compared to state-of-the-art baselines, it reduces transmission payload and GPU VRAM footprint by up to 2.4imes and 5.5imes, respectively, and achieves smooth quality transitions optimal for real-time adaptive streaming. Project page: https://yuang-ian.github.io/evogs/

Source: arXiv cs.CV | 2026-06-08