GaMO: Geometry-aware Multi-view Diffusion Outpainting for Sparse-View 3D Reconstruction
Image generated by Gemini AI
Recent advancements in 3D reconstruction using GaMO (Geometry-aware Multi-view Outpainter) tackle the limitations of existing methods that struggle with limited input views. By expanding the field of view from current camera poses, GaMO maintains geometric consistency and enhances scene coverage. In tests on Replica and ScanNet++, it achieved superior reconstruction quality and a $25\times$ speedup over leading diffusion methods, processing within 10 minutes. For more details, visit the project page: https://yichuanh.github.io/GaMO/.
GaMO Revolutionizes 3D Reconstruction with Geometry-aware Multi-view Diffusion Outpainting
Researchers have unveiled GaMO (Geometry-aware Multi-view Outpainter), a framework that enhances 3D reconstruction from sparse-view inputs. By expanding the existing field of view rather than generating new camera viewpoints, GaMO addresses limitations faced by current methods.
GaMO employs a multi-view outpainting strategy that leverages existing camera poses to enhance scene detail while maintaining geometric consistency. The framework utilizes multi-view conditioning and geometry-aware denoising methods in a zero-shot approach, eliminating the need for prior training.
Extensive evaluations on datasets like Replica and ScanNet++ show that GaMO achieves superior reconstruction quality with 3, 6, and 9 input views, surpassing existing models in Peak Signal-to-Noise Ratio (PSNR) and Learned Perceptual Image Patch Similarity (LPIPS) metrics. It also boasts a $25\times$ improvement in processing speed over state-of-the-art diffusion-based methods, with total processing times under 10 minutes.
For further details, visit the GaMO Project Page.
Related Topics:
📰 Original Source: https://arxiv.org/abs/2512.25073v1
All rights and credit belong to the original publisher.