Research map / 3D vision / Depth and stereo
SCE-Depth: A Spherical Compound Eye Framework for Wide FOV Depth Estimation
Yi Zhu, Hao Xiong, Lin Xiao, Ranfeng Shi, Qinying Gu, Leilei Gu
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 |
|---|---|---|---|
| use native SCE spherical images instead of fisheye images reprojected to HEALPix (ablation) | HealSwin · CompoundDepth (alpha=3 deg) | Abs Rel: 0.106 -> 0.082 | Improved |
| add Spherical Gradient Loss alongside MSE (ablation) | SCE-Depth · CompoundDepth (alpha=3 deg) | RMSE: 0.350 -> 0.281 | Improved |
| add Spherical Gradient Feature Extractor (SGFE) (ablation) | SCE-Depth · CompoundDepth (alpha=3 deg) | RMSE: 0.281 -> 0.256 | Improved |
| use Sobel instead of Laplacian operator in SGFE (ablation) | SCE-Depth · CompoundDepth (alpha=3 deg) | RMSE: 0.295 -> 0.256 | Improved |
| use Sobel instead of Scharr operator in SGFE (ablation) | SCE-Depth · CompoundDepth (alpha=3 deg) | RMSE: 0.258 -> 0.256 | Improved |
| set ommatidial FOV to 3 deg instead of 1 deg (ablation) | SCE-Depth · CompoundDepth | RMSE: 0.279 -> 0.256 | Improved |
| widen ommatidial FOV to 5 deg from 3 deg (ablation) | SCE-Depth · CompoundDepth | RMSE: 0.256 -> 0.289 | Hurt |
| widen ommatidial FOV to 7 deg from 3 deg (ablation) | SCE-Depth · CompoundDepth | RMSE: 0.256 -> 0.284 | Hurt |
| apply SCE-Depth gradient modules on Stanford2D3D versus SUFormer | SCE-Depth · Stanford2D3D | Abs Rel: 0.071 -> 0.114 | Mixed |
| use spherical backbone instead of planar networks on fisheye data | HealSwin vs Swin · FisheyeDepth | RMSE: 0.324 -> 0.361 | Hurt |
Closest papers by meaning
- SynLF: Zero-Shot Metric Depth from Light Field Cameras via Physics-Grounded Synthesis (ECCV 2026 · similarity 0.85)
- Deep Optics for Monocular Depth Estimation and 3D Object Detection (ICCV 2019 · similarity 0.85)
- Toward Real-world BEV Perception: Depth Uncertainty Estimation via Gaussian Splatting (CVPR 2025 · similarity 0.84)
- Distilling Monocular Foundation Model for Fine-grained Depth Completion (CVPR 2025 · similarity 0.84)
- PatchFusion: An End-to-End Tile-Based Framework for High-Resolution Monocular Metric Depth Estimation (CVPR 2024 · similarity 0.83)
- Mind The Edge: Refining Depth Edges in Sparsely-Supervised Monocular Depth Estimation (CVPR 2024 · similarity 0.83)
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.
