arXiv 2026

FreeStreamGS: Online Feed-forward 3D Gaussian Splatting from Unposed Streaming Inputs

TL;DR: A causal feed-forward framework for online 3DGS and NVS from streaming, unposed image inputs.

Ruiyang Chen1 Feiran Li3 Chu Zhou2 Zonglin Li1 Zhanyu Ma1 Heng Guo1,*
1Beijing University of Posts and Telecommunications, China 2National Institute of Informatics, Japan 3Independent Researcher *Corresponding author
FreeStreamGS incrementally reconstructs 3D Gaussian scenes from streaming inputs and renders novel views.

FreeStreamGS incrementally reconstructs renderable 3D Gaussian representations from unposed video streams for online 3DGS and NVS.

Abstract

Feed-forward 3D Gaussian Splatting (3DGS) allows efficient and high-fidelity novel view synthesis (NVS) from an offline recorded image sequence. However, achieving online NVS from streaming and unposed image inputs remains challenging. Although online feed-forward geometric estimation methods have been proposed for streaming depth and point cloud recovery, they cannot be adapted to NVS due to severe rendering artifacts. This is because NVS demands stricter multi-view consistency in Gaussian scales and pose-geometry alignment; even minor deviations would accumulate over time and visibly degrade rendering quality. To this end, we propose FreeStreamGS, a robust online feed-forward framework for efficient and high-quality NVS. We introduce two key mechanisms: a Decoupled Intrinsic Recovery Head that removes cumulative camera intrinsic bias and prevents scene scale jitter during long-term streaming, and a Dynamic Point Refinement Offset strategy that relaxes rigid unprojection to correct coupled pose-depth drift. Extensive experiments show that FreeStreamGS achieves rendering quality competitive with state-of-the-art offline feed-forward 3DGS methods, despite operating without access to future frames.

Method

Overview of the FreeStreamGS pipeline.

The pipeline uses causal feature extraction, decoupled camera recovery, Gaussian decoding with DPR-Offsets, and online recursive Gaussian fusion.

Causal Streaming

Processes frames sequentially with historical context and no future observations.

Stable Intrinsics

Decoupled intrinsic recovery reduces scale jitter during long-term streaming.

Geometry Refinement

DPR-Offsets compensate for coupled pose-depth drift before Gaussian fusion.

Quantitative Results

Dataset Method Online 5-view PSNR 5-view SSIM 5-view LPIPS 64-view PSNR 64-view SSIM 64-view LPIPS
DL3DV-140 WorldMirror No 21.559 0.673 0.285 OOM OOM OOM
DL3DV-140 OnTheFly-NVS Yes 21.030 0.644 0.325 18.478 0.610 0.455
DL3DV-140 FreeStreamGS Yes 21.884 0.688 0.273 17.190 0.558 0.485
RE10K WorldMirror No 25.419 0.820 0.203 OOM OOM OOM
RE10K OnTheFly-NVS Yes 20.829 0.815 0.290 20.985 0.767 0.303
RE10K FreeStreamGS Yes 25.797 0.833 0.200 21.562 0.749 0.297

FreeStreamGS is evaluated under strict online constraints and remains scalable when offline feed-forward baselines run out of memory.

Qualitative Results

GT
FreeStreamGS
OnTheFly-NVS
WorldMirror
AnySplat
FLARE
DL3DV ground truth sample 1 FreeStreamGS DL3DV sample 1 OnTheFly-NVS DL3DV sample 1 WorldMirror DL3DV sample 1 AnySplat DL3DV sample 1 FLARE DL3DV sample 1
DL3DV ground truth sample 2 FreeStreamGS DL3DV sample 2 OnTheFly-NVS DL3DV sample 2 WorldMirror DL3DV sample 2 AnySplat DL3DV sample 2 FLARE DL3DV sample 2
RE10K ground truth sample 1 FreeStreamGS RE10K sample 1 OnTheFly-NVS RE10K sample 1 WorldMirror RE10K sample 1 AnySplat RE10K sample 1 FLARE RE10K sample 1
RE10K ground truth sample 2 FreeStreamGS RE10K sample 2 OnTheFly-NVS RE10K sample 2 WorldMirror RE10K sample 2 AnySplat RE10K sample 2 FLARE RE10K sample 2

Qualitative comparison on DL3DV-140 and RE10K. FreeStreamGS preserves structural detail while operating causally.

Video Stabilization

FreeStreamGS video stabilization result 15 FreeStreamGS video stabilization result 21 FreeStreamGS video stabilization result 25

Once a causal Gaussian representation is available, FreeStreamGS can render stabilized views along smoothed camera trajectories.

BibTeX

@article{chen2026freestreamgs,
  title   = {FreeStreamGS: Online Feed-forward 3D Gaussian Splatting from Unposed Streaming Inputs},
  author  = {Chen, Ruiyang and Li, Feiran and Zhou, Chu and Li, Zonglin and Ma, Zhanyu and Guo, Heng},
  journal = {arXiv preprint arXiv:2606.03254},
  year    = {2026}
}