Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.02102 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908748298059776 |
|---|---|
| author | Yao, Jiaqi Yan, Zhongmiao Xu, Jingyi Xia, Songpengcheng Xiang, Yan Pei, Ling |
| author_facet | Yao, Jiaqi Yan, Zhongmiao Xu, Jingyi Xia, Songpengcheng Xiang, Yan Pei, Ling |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_02102 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | 360-GeoGS: Geometrically Consistent Feed-Forward 3D Gaussian Splatting Reconstruction for 360 Images Yao, Jiaqi Yan, Zhongmiao Xu, Jingyi Xia, Songpengcheng Xiang, Yan Pei, Ling Computer Vision and Pattern Recognition 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. |
| title | 360-GeoGS: Geometrically Consistent Feed-Forward 3D Gaussian Splatting Reconstruction for 360 Images |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.02102 |