DUSt3R: Geometric 3D Vision Made Easy

A feed-forward network that regresses two dense pointmaps in a shared coordinate frame from an uncalibrated, unposed image pair, jointly recovering correspondence, relative pose, intrinsics and depth without prior calibration.

Recovering 3D geometry — point maps, depth, and camera poses — directly from images in a single network pass, replacing the detect-match-triangulate-bundle-adjust pipeline of classical structure-from-motion with learned pointmap regression.