Evaluating Robustness of Monocular Depth Estimation with Procedural Scene Perturbations
This work addresses the need for better robustness evaluation in monocular depth estimation, which is incremental as it builds on existing benchmarks by adding systematic testing.
The authors tackled the problem of incomplete robustness assessment in monocular depth estimation by introducing PDE, a new benchmark using procedural generation to test robustness to controlled perturbations like object, camera, material, and lighting changes, revealing challenging perturbations for state-of-the-art models.
Recent years have witnessed substantial progress on monocular depth estimation, particularly as measured by the success of large models on standard benchmarks. However, performance on standard benchmarks does not offer a complete assessment, because most evaluate accuracy but not robustness. In this work, we introduce PDE (Procedural Depth Evaluation), a new benchmark which enables systematic robustness evaluation. PDE uses procedural generation to create 3D scenes that test robustness to various controlled perturbations, including object, camera, material and lighting changes. Our analysis yields interesting findings on what perturbations are challenging for state-of-the-art depth models, which we hope will inform further research. Code and data are available at https://github.com/princeton-vl/proc-depth-eval.