Research map / 3D vision / Depth and stereo
OmniDS: Dual-Stream Context Fusion for Omnidirectional Depth from Fisheye Cameras
Chaesong Park, Jihyeon Hwang, Muyeol Sung, Jongwoo Lim
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 both CNN and DINO as context instead of CNN only (ablation) | OmniDS · Sunny ablation | MAE: 0.38 -> 0.35 | Improved |
| use both CNN and DINO context instead of CNN only (ablation) | OmniDS · Sunny ablation | >1 error: 5.06 -> 4.69 | Improved |
| use CNN similarity instead of DINO similarity (ablation) | OmniDS · Sunny ablation | MAE: 0.51 -> 0.38 | Improved |
| add multi-view consensus volume (MVC) to CNN+DINO context (ablation) | OmniDS · Sunny ablation | >1 error: 4.69 -> 4.60 | Improved |
| add MVC to CNN+DINO context (ablation) | OmniDS · Sunny ablation | RMS: 1.55 -> 1.63 | Hurt |
| add deformable cross-attention with distortion bias (ablation) | OmniDS · Sunny ablation | >1 error: 4.60 -> 4.49 | Improved |
| add 2D spatial aggregation to correlation profile (ablation) | OmniDS · Sunny ablation (similarity volumes) | >1 error: 5.00 -> 4.67 | Improved |
| add MVC to Corr + 2D Aggr. (ablation) | OmniDS · Sunny ablation (similarity volumes) | >1 error: 4.67 -> 4.49 | Improved |
| add MVC to Corr + 2D Aggr. (ablation) | OmniDS · Sunny ablation (similarity volumes) | RMS: 1.52 -> 1.59 | Hurt |
| distill dual-stream encoder into MobileNet student | OmniDS Ours-ft · OmniThings (fine-tuned) | MAE: 1.70 -> 2.11 | Hurt |
| distill dual-stream encoder into MobileNet student | OmniDS Ours-ft · inference time | time (ms): 148 -> 102 | Improved |
| distill dual-stream encoder into MobileNet student | OmniDS Ours-ft · OmniHouse (fine-tuned) | >1 error: 3.92 -> 3.41 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MDP-Omni — MDP-Omni: Parameter-free Multimodal Depth Prior-based Sampling for Omnidirectional Stereo Matching (ICCV 2025) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MDP-Omni: Parameter-free Multimodal Depth Prior-based Sampling for Omnidirectional Stereo Matching (ICCV 2025 · similarity 0.89)
- Extending Foundational Monocular Depth Estimators to Fisheye Cameras with Calibration Tokens (ICCV 2025 · similarity 0.89)
- OmniMVS: End-to-End Learning for Omnidirectional Stereo Matching (ICCV 2019 · similarity 0.89)
- OmniStereo: Real-time Omnidireactional Depth Estimation with Multiview Fisheye Cameras (CVPR 2025 · similarity 0.86)
- HELVIPAD: A Real-World Dataset for Omnidirectional Stereo Depth Estimation (CVPR 2025 · similarity 0.85)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.85)
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.
