Depth Anything in $360^\circ$: Towards Scale Invariance in the Wild

TL;DR

This paper introduces DA360, a panoramic depth estimation model that achieves scale invariance and seamless spherical depth maps, significantly improving zero-shot outdoor and indoor depth estimation performance.

Contribution

We propose a novel scale-invariance learning approach for panoramic depth estimation and integrate circular padding to improve spatial coherence, advancing zero-shot generalization in open-world environments.

Findings

Over 50% error reduction on indoor benchmarks

Over 10% error reduction on outdoor datasets

30% improvement over PanDA in zero-shot panoramic depth estimation

Abstract

Panoramic depth estimation provides a comprehensive solution for capturing complete $360^\circ$ environmental structural information, offering significant benefits for robotics and AR/VR applications. However, while extensively studied in indoor settings, its zero-shot generalization to open-world domains lags far behind perspective images, which benefit from abundant training data. This disparity makes transferring capabilities from the perspective domain an attractive solution. To bridge this gap, we present Depth Anything in $360^\circ$ (DA360), a panoramic-adapted version of Depth Anything V2. Our key innovation involves learning a shift parameter from the ViT backbone, transforming the model's scale- and shift-invariant output into a scale-invariant estimate that directly yields well-formed 3D point clouds. This is complemented by integrating circular padding into the DPT decoder to eliminate seam artifacts, ensuring spatially coherent depth maps that respect spherical continuity. Evaluated on standard indoor benchmarks and our newly curated outdoor dataset, Metropolis, DA360 shows substantial gains over its base model, achieving over 50\% and 10\% relative depth error reduction on indoor and outdoor benchmarks, respectively. Furthermore, DA360 significantly outperforms robust panoramic depth estimation methods, achieving about 30\% relative error improvement compared to PanDA across all three test datasets and establishing new state-of-the-art performance for zero-shot panoramic depth estimation.