Stereo-GS: Multi-View Stereo Vision Model for Generalizable 3D Gaussian Splatting Reconstruction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Xiufeng, Cheung, Ka Chun, Cong, Runmin, See, Simon, Wan, Renjie
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915702588309504
author Huang, Xiufeng
Cheung, Ka Chun
Cong, Runmin
See, Simon
Wan, Renjie
author_facet Huang, Xiufeng
Cheung, Ka Chun
Cong, Runmin
See, Simon
Wan, Renjie
contents Generalizable 3D Gaussian Splatting reconstruction showcases advanced Image-to-3D content creation but requires substantial computational resources and large datasets, posing challenges to training models from scratch. Current methods usually entangle the prediction of 3D Gaussian geometry and appearance, which rely heavily on data-driven priors and result in slow regression speeds. To address this, we propose \method, a disentangled framework for efficient 3D Gaussian prediction. Our method extracts features from local image pairs using a stereo vision backbone and fuses them via global attention blocks. Dedicated point and Gaussian prediction heads generate multi-view point-maps for geometry and Gaussian features for appearance, combined as GS-maps to represent the 3DGS object. A refinement network enhances these GS-maps for high-quality reconstruction. Unlike existing methods that depend on camera parameters, our approach achieves pose-free 3D reconstruction, improving robustness and practicality. By reducing resource demands while maintaining high-quality outputs, \method provides an efficient, scalable solution for real-world 3D content generation.
format Preprint
id arxiv_https___arxiv_org_abs_2507_14921
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Stereo-GS: Multi-View Stereo Vision Model for Generalizable 3D Gaussian Splatting Reconstruction
Huang, Xiufeng
Cheung, Ka Chun
Cong, Runmin
See, Simon
Wan, Renjie
Computer Vision and Pattern Recognition
Generalizable 3D Gaussian Splatting reconstruction showcases advanced Image-to-3D content creation but requires substantial computational resources and large datasets, posing challenges to training models from scratch. Current methods usually entangle the prediction of 3D Gaussian geometry and appearance, which rely heavily on data-driven priors and result in slow regression speeds. To address this, we propose \method, a disentangled framework for efficient 3D Gaussian prediction. Our method extracts features from local image pairs using a stereo vision backbone and fuses them via global attention blocks. Dedicated point and Gaussian prediction heads generate multi-view point-maps for geometry and Gaussian features for appearance, combined as GS-maps to represent the 3DGS object. A refinement network enhances these GS-maps for high-quality reconstruction. Unlike existing methods that depend on camera parameters, our approach achieves pose-free 3D reconstruction, improving robustness and practicality. By reducing resource demands while maintaining high-quality outputs, \method provides an efficient, scalable solution for real-world 3D content generation.
title Stereo-GS: Multi-View Stereo Vision Model for Generalizable 3D Gaussian Splatting Reconstruction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.14921