[1/4] Camera poses are essential for (neural) 3D reconstruction. But what about sparse-view settings where obtaining these via COLMAP isn’t feasible?

[1/4] Camera poses are essential for (neural) 3D reconstruction. But what about sparse-view settings where obtaining these via COLMAP isn’t feasible?