Multi-View Geometry

Follow this topic Ask about this topic

Multi-view geometry studies how to infer camera relationships and three-dimensional structure from images captured at different viewpoints, enabling mapping, measurement, tracking, and reconstruction. Classical methods use feature correspondences, epipolar constraints, homographies, triangulation, and bundle adjustment, while learned matchers and geometry-aware transformers extend these tools to difficult images and larger view sets. Current work combines explicit geometric constraints with learned depth and foundation-model features across settings such as wide-baseline, uncalibrated, fisheye, aerial, and satellite imagery. A central challenge is making estimates reliable under weak overlap, noise, motion, and calibration errors while keeping methods efficient and evaluating accuracy in ways that detect coherent geometric failures.

Latest papers

97 papers10 papers in the last four weeks, up 25% on the four weeks before. 0.1% of all new papers.

GGPT: Geometry Grounded Point Transformer
Mar 11, 2026Yutong Chen, Yiming Wang, Xucong Zhang +2Sparse-View 3D Reconstruction Feed-Forward 3D Reconstruction
ETH Zurich · Delft University of Technology
GloSplat: Joint Pose-Appearance Optimization for Faster and More Accurate 3D Reconstruction
Mar 5, 2026Tianyu Xiong, Rui Li, Linjie Li +13D Reconstruction 3D Gaussian Splatting
Department of Computer Science, Northwestern Polytechnical University, Shanxi, China · CEMSE, KAUST, Thuwal, Jeddah
Affine Correspondences in Stereo Vision: Theory, Practice, and Limitations
Mar 2, 2026Levente HajderStereo Vision Multi-View Geometry
Faculty of Informatics, Eötvös Loránd University, Pázmány Péter stny. 1/C, Budapest, H-1117, Hungary
Geometry OR Tracker: Universal Geometric Operating Room Tracking
Feb 28, 2026Yihua Shao, Kang Chen, Feng Xue +6Point Tracking Multi-View Geometry
Centre for Artificial Intelligence and Robotics Hong Kong Institute of Science and Innovation Chinese Academy of Sciences · State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS) Institute of Automation Chinese Academy of Sciences (CASIA) · Peking University +4
Multiprojective Geometry of Compatible Triples of Fundamental and Essential Matrices
Feb 26, 2026Timothy Duff, Viktor Korotynskiy, Anton Leykin +1Multi-View Geometry
QuadSync: Quadrifocal Tensor Synchronization via Tucker Decomposition
Feb 26, 2026Daniel Miao, Gilad Lerman, Joe KileelMulti-View Geometry Structure from Motion
School of Mathematics, University of Minnesota · Department of Mathematics and Oden Institute for Computational Engineering and Sciences, University of Texas at Austin
Segmentation-Guided Homography Estimation for Long-Term Planar Tracking
Feb 23, 2026Jonas Serych, Jiri MatasVisual Object Tracking Multi-View Geometry
Visual Recognition Group, Faculty of Electrical Engineering, Czech Technical University in Prague, Czech Republic
Image Matching Filtering and Refinement by Planes and Beyond
Nov 14, 2024Fabio Bellavia, Zhenjun Zhao, Luca Morelli +1Multi-View Geometry Image Matching
University of Palermo, Italy · University of Zaragoza, Spain · Bruno Kessler Foundation, Italy

Mar 11, 2026J/K move · Enter open · S save

GGPT: Geometry Grounded Point Transformer

Yutong Chen, Yiming Wang, Xucong Zhang, Sergey Prokudin+1

ETH Zurich · Delft University of Technology

Sparse-View 3D Reconstruction Feed-Forward 3D Reconstruction 3D ViTs Multi-View Geometry

Open paper AskPDF arXiv

Recent feed-forward networks have achieved remarkable progress in sparse-view 3D reconstruction by predicting dense point maps directly from RGB images. However, they often suffer from geometric inconsistencies and limited fine-grained accuracy due to the absence of explicit multi-view constraints. We introduce the Geometry-Grounded Point Transformer (GGPT), a framework that augments feed-forward reconstruction with reliable sparse geometric guidance. We first propose an improved Structure-from-Motion pipeline based on dense feature matching and lightweight geometric optimisation to efficiently estimate accurate camera poses and partial 3D point clouds from sparse input views. Building on this foundation, we propose a geometry-guided 3D point transformer that refines dense point maps under explicit partial-geometry supervision using an optimised guidance encoding. Extensive experiments demonstrate that our method provides a principled mechanism for integrating geometric priors with dense feed-forward predictions, producing reconstructions that are both geometrically consistent and spatially complete, recovering fine structures and filling gaps in textureless areas. Trained solely on ScanNet++ with VGGT predictions, GGPT generalises across architectures and datasets, substantially outperforming state-of-the-art feed-forward 3D reconstruction models in both in-domain and out-of-domain settings.