Research map / 3D vision / Depth and stereo
DualRefine: Self-Supervised Depth and Pose Estimation Through Iterative Epipolar Sampling and Refinement Toward Equilibrium
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| iteratively refine depth and pose together using local epipolar matching | DualRefine · KITTI Eigen split | Abs Rel: 0.103 -> 0.087 | Helped |
| iteratively refine depth and pose together | DualRefine · KITTI odometry Seq. 09 | ATE (m): 39.31 -> 5.18 | Helped |
| iteratively refine depth and pose together | DualRefine · KITTI odometry Seq. 10 | ATE (m): 15.13 -> 10.85 | Helped |
| add pose updates (ablation) | DualRefine · KITTI Eigen split | Abs Rel: 0.097 -> 0.090 | Helped |
| add evolving confidence weights Wh,k (ablation) | DualRefine pose updates · KITTI Eigen split | Abs Rel: 0.090 -> 0.091 | No clear effect |
| add input-context confidence weights Wq (ablation) | DualRefine pose updates · KITTI Eigen split | Abs Rel: 0.088 -> 0.087 | No clear effect |
| use refined pose T* rather than initial pose T0 for the consistency mask (ablation) | DualRefine · KITTI Eigen split | Abs Rel: 0.090 -> 0.087 | Helped |
| use 6 rather than 3 test iterations with 6 training iterations (ablation) | DualRefine DEQ · KITTI Eigen split | Abs Rel: 0.097 -> 0.087 | Helped |
| increase training and test iterations from 6 to 12 (ablation) | DualRefine DEQ · KITTI Eigen split | Abs Rel: 0.087 -> 0.089 | Hurt |
| add iterative refinement for 6 steps (ablation) | DualRefine · KITTI Eigen split | Abs Rel: 0.104 -> 0.087 | Helped |
| pair refined depth and pose estimates for the self-supervision losses (ablation) | DualRefine · KITTI Eigen split | Abs Rel: 0.093 -> 0.089 | Helped |
Closest papers by meaning
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.88)
- Relative Pose Estimation through Affine Corrections of Monocular Depth Priors (CVPR 2025 · similarity 0.88)
- GeoDepth: From Point-to-Depth to Plane-to-Depth Modeling for Self-Supervised Monocular Depth Estimation (CVPR 2025 · similarity 0.86)
- DualPM: Dual Posed-Canonical Point Maps for 3D Shape and Pose Reconstruction (CVPR 2025 · similarity 0.86)
- SQLdepth: Generalizable Self-Supervised Fine-Structured Monocular Depth Estimation (AAAI 2024 · similarity 0.86)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · 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.