Research map / 3D vision / Depth and stereo
Extending Foundational Monocular Depth Estimators to Fisheye Cameras with Calibration Tokens
Suchisrit Gangopadhyay, Jung-Hee Kim, Xien Chen, Patrick Rim, Hyoungseob Park, Alex Wong
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add layer-wise Calibration Tokens (unique per encoder layer) instead of a single token set (ablation) | UniDepth + Calibration Tokens · ScanNet++ | RMSE: 0.254 -> 0.244 | Improved |
| add layer-wise Calibration Tokens instead of a single token set (ablation) | UniDepth + Calibration Tokens · KITTI-360 | RMSE: 2.065 -> 2.040 | Improved |
| use LogL1 loss instead of L1 (ablation) | UniDepth + Calibration Tokens · ScanNet++ | RMSE: 0.260 -> 0.254 | Improved |
| compute loss in perspective frame (inverse-warp fisheye output) instead of fisheye frame (ablation) | UniDepth + Calibration Tokens · ScanNet++ | RMSE: 0.280 -> 0.244 | Improved |
| use unique tokens per layer rather than one same token re-added every block (ablation) | UniDepth + Calibration Tokens · ScanNet++ | RMSE: 0.290 -> 0.244 | Improved |
| train tokens with ground-truth depth rather than self-supervised FMDE pseudo-labels (ablation) | UniDepth + Calibration Tokens · KITTI-360 | RMSE: 2.040 -> 1.994 | Improved |
| use Calibration Tokens instead of full finetuning on synthetic fisheye images | MiDaS · KITTI-360 | RMSE: 9.289 -> 2.348 | Improved |
| add Calibration Tokens to DepthAnything | DepthAnything · KITTI-360 | delta1: 0.839 -> 0.810 | Mixed |
| add Calibration Tokens to UniDepth | UniDepth · KITTI-360 | delta1: 0.663 -> 0.664 | No clear effect |
| add Calibration Tokens to UniDepth | UniDepth · ScanNet++ | RMSE: 0.279 -> 0.244 | Improved |
Closest papers by meaning
- From Perspective to Fisheye Depth Estimation and Open-Vocabulary Segmentation (ECCV 2026 · similarity 0.93)
- DepthMaster: Unified Monocular Depth Estimation for Perspective and Panoramic Images (NeurIPS 2026 · similarity 0.86)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.86)
- Relative Pose Estimation through Affine Corrections of Monocular Depth Priors (CVPR 2025 · similarity 0.86)
- Deep Single Image Camera Calibration by Heatmap Regression to Recover Fisheye Images Under Manhattan World Assumption (CVPR 2024 · similarity 0.86)
- UniSHARP: Universal Sharp Monocular View Synthesis (NeurIPS 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
