360-GeoGS: Geometrically Consistent Feed-Forward 3D Gaussian Splatting Reconstruction for 360 Images

2026-01-05

arXiv.org

AI Summary

Existing feedforward 3D Gaussian splatting methods suffer from insufficient geometric consistency in 360° image reconstruction, limiting their applicability to high-precision spatial perception tasks. This work proposes a feedforward 3D Gaussian splatting framework tailored for 360° imagery, which introduces a depth-normal coupled geometric regularization mechanism to jointly optimize the position, rotation, and scale of Gaussian ellipsoids. By doing so, the method significantly enhances the geometric consistency of both point cloud and surface reconstructions while preserving high-fidelity rendering quality. To the best of our knowledge, this is the first approach within the feedforward Gaussian splatting paradigm to achieve simultaneous optimization of rendering fidelity and geometric accuracy, thereby establishing a reliable foundation for applications such as augmented reality, robotics, and digital twins.

Technology Category

• Computer Vision: 3D Computer Vision
• Search and Optimization: Sampling/Simulation-based Search
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

3D scene reconstruction is fundamental for spatial intelligence applications such as AR, robotics, and digital twins. Traditional multi-view stereo struggles with sparse viewpoints or low-texture regions, while neural rendering approaches, though capable of producing high-quality results, require per-scene optimization and lack real-time efficiency. Explicit 3D Gaussian Splatting (3DGS) enables efficient rendering, but most feed-forward variants focus on visual quality rather than geometric consistency, limiting accurate surface reconstruction and overall reliability in spatial perception tasks. This paper presents a novel feed-forward 3DGS framework for 360 images, capable of generating geometrically consistent Gaussian primitives while maintaining high rendering quality. A Depth-Normal geometric regularization is introduced to couple rendered depth gradients with normal information, supervising Gaussian rotation, scale, and position to improve point cloud and surface accuracy. Experimental results show that the proposed method maintains high rendering quality while significantly improving geometric consistency, providing an effective solution for 3D reconstruction in spatial perception tasks.

Problem

• 3D reconstruction
• geometric consistency
• 3D Gaussian Splatting
• 360 images
• spatial perception

Innovation

• 3D Gaussian Splatting
• geometric consistency
• Depth-Normal regularization
• feed-forward reconstruction
• 360 image