Bi-FlowGS: Bridging Generative View Completion and Gaussian Geometry through Bidirectional Flow Co-Refinement

arXiv:2609.17039v1 Announce Type: new Abstract: Sparse-view 3D scene reconstruction with 3D Gaussian Splatting (3DGS) is inherently underconstrained. Plausible renderings can also coexist with erroneous Gaussian geometry, as errors in positions or depths may be concealed by opacity, scale, and appearance; we term this failure mode Geometry Cheating. Existing regularization methods constrain geometry but remain limited to observed views, while video-diffusion-based methods complete unseen views yet mainly use them as RGB pseudo-supervision, underusing motion and temporal priors and lacking explicit geometry supervision. We present Bi-FlowGS, which uses optical flow to bridge generative view completion and Gaussian geometry regularization. Our plug-and-play Video-to-Geometry Flow Distillation (V2G) distills temporal correspondence priors from restored videos into Gaussian geometry to alleviate Geometry Cheating. Conversely, Geometry-to-Video Flow-Guided Restoration (G2V) uses the current 3DGS geometry to guide temporally consistent video restoration, providing more reliable generative supervision. Together, V2G and G2V form an implicit bidirectional co-refinement process, enabling restored videos and the optimized 3DGS scene to iteratively improve each other. Experiments demonstrate improved rendering quality and geometric consistency across wide-baseline and unbounded 360{eg} benchmarks.

Related

• SatSplatDiff: Geometry-preserving generative refinement for high-fidelity satellite Gaussian Splatting
• GS-ROR^2: Bidirectional-guided 3DGS and SDF for Reflective Object Relighting and Reconstruction
• From Uncertainty to Stability and Fidelity: Guiding Sparse-View 3D Gaussian Splatting with Fisher Information
• ConFixGS: Learning to Fix Feedforward 3D Gaussian Splatting with Confidence-Aware Diffusion Priors in Driving Scenes

Source: arXiv cs.CV | 2026-09-16