MapAnything: Universal Feed-Forward Metric 3D Reconstruction
By: Nikhil Keetha , Norman Müller , Johannes Schönberger and more
Potential Business Impact:
Builds 3D worlds from pictures and camera data.
We introduce MapAnything, a unified transformer-based feed-forward model that ingests one or more images along with optional geometric inputs such as camera intrinsics, poses, depth, or partial reconstructions, and then directly regresses the metric 3D scene geometry and cameras. MapAnything leverages a factored representation of multi-view scene geometry, i.e., a collection of depth maps, local ray maps, camera poses, and a metric scale factor that effectively upgrades local reconstructions into a globally consistent metric frame. Standardizing the supervision and training across diverse datasets, along with flexible input augmentation, enables MapAnything to address a broad range of 3D vision tasks in a single feed-forward pass, including uncalibrated structure-from-motion, calibrated multi-view stereo, monocular depth estimation, camera localization, depth completion, and more. We provide extensive experimental analyses and model ablations demonstrating that MapAnything outperforms or matches specialist feed-forward models while offering more efficient joint training behavior, thus paving the way toward a universal 3D reconstruction backbone.
Similar Papers
MapAnything: Universal Feed-Forward Metric 3D Reconstruction
CV and Pattern Recognition
Builds 3D worlds from pictures and camera data.
MapAnything: Mapping Urban Assets using Single Street-View Images
CV and Pattern Recognition
Finds city objects and problems from pictures.
Layout Anything: One Transformer for Universal Room Layout Estimation
CV and Pattern Recognition
Draws accurate room layouts from pictures fast.