10 upvotes
Looped Transformer Turns Multi-View Images into 3D Maps
What is Solv3R: 3D Looped Transformer as a Multi-view Solver?
Solv3R is a transformer-based model that iteratively refines camera poses and scene geometry from a set of overlapping photos. It mimics classic structure-from-motion while learning the update steps, allowing new views to be added without recomputing the whole reconstruction.
Who it is for
Computer-vision researchers and AR/VR developers can use it to speed up and learn 3D reconstructions from multi-view data.