PolarGuide-GSDR: 3D Gaussian Splatting Driven by Polarization Priors and Deferred Reflection for Real-World Reflective Scenes

2025-12-02

AI Summary

Existing polarization-aware NeRF methods suffer from slow training, inefficient rendering, and reliance on strong material or viewpoint assumptions in specular scenes; while 3D Gaussian Splatting (3DGS) enables real-time rendering, its entanglement of reflection and geometry impedes accurate specular modeling, and existing delayed-reflection modules require environment maps. This paper proposes a polarization-prior-guided 3DGS framework that, for the first time, directly incorporates polarimetric observations into the 3DGS optimization pipeline, establishing a bidirectional coupling between geometry and polarization—enabling high-fidelity decoupling of reflection and geometry without environment maps or explicit material priors. Our method integrates polarimetric imaging, normal optimization, and spherical harmonic-based reflectance representation, leveraging geometric priors to resolve polarization ambiguities and using optimized polarization cues to guide reflection modeling. Evaluated on public and custom datasets, it achieves state-of-the-art performance in specular reconstruction, normal estimation, and novel-view synthesis, while preserving real-time rendering capability.

Technology Category

• Computer Vision: 3D Computer Vision
• Search and Optimization: Non-convex Optimization
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT
• Security and Privacy: Data transparency and provenance

Abstract

Polarization-aware Neural Radiance Fields (NeRF) enable novel view synthesis of specular-reflection scenes but face challenges in slow training, inefficient rendering, and strong dependencies on material/viewpoint assumptions. However, 3D Gaussian Splatting (3DGS) enables real-time rendering yet struggles with accurate reflection reconstruction from reflection-geometry entanglement, adding a deferred reflection module introduces environment map dependence. We address these limitations by proposing PolarGuide-GSDR, a polarization-forward-guided paradigm establishing a bidirectional coupling mechanism between polarization and 3DGS: first 3DGS's geometric priors are leveraged to resolve polarization ambiguity, and then the refined polarization information cues are used to guide 3DGS's normal and spherical harmonic representation. This process achieves high-fidelity reflection separation and full-scene reconstruction without requiring environment maps or restrictive material assumptions. We demonstrate on public and self-collected datasets that PolarGuide-GSDR achieves state-of-the-art performance in specular reconstruction, normal estimation, and novel view synthesis, all while maintaining real-time rendering capabilities. To our knowledge, this is the first framework embedding polarization priors directly into 3DGS optimization, yielding superior interpretability and real-time performance for modeling complex reflective scenes.

Problem

• Reconstructs reflective scenes accurately without environment maps.
• Separates reflections from geometry using polarization-guided 3D Gaussian Splatting.
• Achieves real-time rendering and high-fidelity novel view synthesis.

Innovation

• Polarization guides 3D Gaussian Splatting for reflection separation
• Bidirectional coupling resolves polarization ambiguity using geometric priors
• Eliminates environment maps and material assumptions for real-time rendering