J
Jan-Michael Frahm
Researcher at University of North Carolina at Chapel Hill
Publications - 200
Citations - 16425
Jan-Michael Frahm is an academic researcher from University of North Carolina at Chapel Hill. The author has contributed to research in topics: Structure from motion & 3D reconstruction. The author has an hindex of 52, co-authored 193 publications receiving 12967 citations. Previous affiliations of Jan-Michael Frahm include University of Kiel & Facebook.
Papers
More filters
Proceedings ArticleDOI
Structure-from-Motion Revisited
TL;DR: This work proposes a new SfM technique that improves upon the state of the art to make a further step towards building a truly general-purpose pipeline.
Book ChapterDOI
Pixelwise View Selection for Unstructured Multi-View Stereo
TL;DR: The core contributions are the joint estimation of depth andnormal information, pixelwise view selection using photometric and geometric priors, and a multi-view geometric consistency term for the simultaneous refinement and image-based depth and normal fusion.
Journal ArticleDOI
Detailed Real-Time Urban 3D Reconstruction from Video
Marc Pollefeys,David Nister,Jan-Michael Frahm,Amir Akbarzadeh,Philippos Mordohai,Brian Clipp,Chris Engels,David Gallup,Seon Joo Kim,Paul Merrell,C. Salmi,Sudipta N. Sinha,B. Talton,Wang Liang,Qingxiong Yang,Henrik Stewenius,Ruigang Yang,Gregory F. Welch,Herman Towles +18 more
TL;DR: A system for automatic, geo-registered, real-time 3D reconstruction from video of urban scenes that extends existing algorithms to meet the robustness and variability necessary to operate out of the lab and shows results on real video sequences comprising hundreds of thousands of frames.
Book ChapterDOI
Building Rome on a cloudless day
Jan-Michael Frahm,Pierre Fite-Georgel,David Gallup,Timothy A. Johnson,Rahul Raguram,Changchang Wu,Yi-Hung Jen,Enrique Dunn,Brian Clipp,Svetlana Lazebnik,Marc Pollefeys +10 more
TL;DR: This paper introduces an approach for dense 3D reconstruction from unregistered Internet-scale photo collections with about 3 million images within the span of a day on a single PC ("cloudless"), leveraging geometric and appearance constraints to obtain a highly parallel implementation on modern graphics processors and multi-core architectures.
Book ChapterDOI
A Comparative Analysis of RANSAC Techniques Leading to Adaptive Real-Time Random Sample Consensus
TL;DR: The technique developed is capable of efficiently adapting to the constraints presented by a fixed time budget, while at the same time providing accurate estimation over a wide range of inlier ratios, and shows significant improvements in accuracy and speed over existing techniques.