$R^3$: 3D Reconstruction via Relative Regression

ArXi:2605.26519v1 Announce Type: new Recent feed-forward geometry foundation models have nstrated impressive generalization by recovering depth and poses in a single forward pass. However, these models are typically constrained by a global coordinate frame assumption. This dependency becomes a significant bottleneck for long-context and streaming reconstruction, as it forces the network to maintain an arbitrary temporal origin and handle translation magnitudes that grow unbounded over time. Our solution, which we call $R^3$, employs relative regression.