In addition, we consider two large-scale models from the DepthAnything<\/strong> family. DepthAnything V1<\/strong> [30] leverages large collections of unlabeled images to improve robustness, while DepthAnything V2<\/strong> [31] follows a two-stage training strategy\u2014pretraining on 500K synthetic images followed by large-scale adaptation on 62 million real-world samples\u2014yielding strong generalization across domains.<\/p>\n\n\n\n Our system is intentionally designed with modularity in mind: new depth estimation methods can be integrated through a common interface with minimal effort, enabling systematic comparison across architectures, data sources, and training objectives.<\/p>\n\n\n\n Our preliminary results will cover some of the results from random sampling on both the 3D-FRONT dataset as well as our own. The \ud835\udeff\u2081 and AbsRel metrics are evaluated for Depth Anything v2 across different poses, with the graph displaying the mean and standard deviation at each pose under randomly sampled rotations. Notably, the \ud835\udeff\u2081 score drops by approximately 25% relative to its average value, while the AbsRel score experiences a similar decline of about 29%, indicating a significant degradation in depth estimation performance with rotational variation.<\/p>\n\n\n\n In addition to the camera parameterization adopted in our main methodology, we conduct an ablation study to analyze the impact of different rotation representations on optimization stability and model performance. Specifically, we compare our chosen parameterization against alternatives based on quaternions<\/strong> and Lie algebra <\/strong> representations.<\/p>\n\n\n\n These parameterizations differ in their geometric properties and optimization behavior. Quaternions provide a compact and continuous representation of 3D rotations but require explicit normalization to maintain unit length during optimization, which can introduce additional constraints. Lie algebra representations, while minimal and well-suited for incremental updates, may exhibit local linearization effects that influence convergence when large rotations are present. By contrast, our selected parameterization offers a smoother optimization landscape for the update process used in our framework.<\/p>\n\n\n\n Through this ablation, we systematically evaluate convergence behavior, numerical stability, and final task performance across these representations, allowing us to isolate the effect of camera parameterization from other modeling choices. This analysis provides insight into how rotation representation influences both optimization dynamics and downstream depth model failure discovery.<\/p>\n","protected":false},"excerpt":{"rendered":" Models Studied To ensure broad and representative coverage of modern monocular depth estimation approaches, our evaluation pipeline incorporates several high-performing models spanning different training regimes and supervision strategies. We include MiDaS [22], a widely adopted method trained across heterogeneous datasets; among its available configurations, we select the highest-capacity variant to maximize performance. We also evaluate … <\/p>\nOther Experiements<\/h2>\n\n\n\n
Random Sampling<\/h2>\n\n\n\n
![\"\"](\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-2\/wp-content\/uploads\/sites\/125\/2025\/05\/Screenshot-2025-05-06-at-2.09.30\u202fPM-1024x956.png\")
<\/figure>\n\n\n\nAblation on Camera Parameterization<\/h2>\n\n\n\n