Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer

A monocular depth network trained for zero-shot cross-dataset transfer by mixing incompatible depth datasets under a scale-and-shift-invariant loss, predicting relative inverse depth up to an unknown global scale and shift.

Predicting per-pixel scene depth from a single image — the scale ambiguity that forces relative (affine-invariant) versus metric formulations, the scale-and-shift-invariant training that lets incompatible datasets be mixed, and the foundation-model recipe that scaled it to zero-shot generalization.