DeepBA: a 2M-parameter GNN learns bundle adjustment and beats prior learned SfM from a cold start

Singularitarian · x · 2026-11-18

A new preprint from TUM researchers presents Deep Bundle Adjustment (DeepBA), a graph neural network with only 2.04M parameters that recovers camera poses, intrinsics, and scene structure from 2D matches across an arbitrary number of views.

Related event: TUM's DeepBA: 2M-Parameter GNN Achieves Cold-Start Bundle Adjustment(2 posts)→

Original post →

More from Research

Research channel →