Research map / 3D vision / Depth and stereo
Learning Accurate Dense Correspondences and When To Trust Them
Check 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 |
|---|---|---|---|
| model flow with a constrained Laplace mixture rather than a single Laplace (ablation) | PDC-Net-s · MegaDepth | PCK-1 (%): 27.45 -> 32.51 | Helped |
| constrain mixture component variance ranges rather than use an unconstrained mixture (ablation) | PDC-Net-s · KITTI-2015 | AUSE: 0.670 -> 0.205 | Helped |
| use a constrained rather than unconstrained Laplace mixture (ablation) | PDC-Net-s · YFCC100M | mAP @5°: 31.18 -> 33.77 | Helped |
| use only the correlation uncertainty module, without integrating flow information (ablation) | PDC-Net-s · KITTI-2015 | AUSE: 0.205 -> 0.418 | Hurt |
| use only the correlation uncertainty module instead of the full uncertainty decoder (ablation) | PDC-Net-s · YFCC100M | mAP @5°: 33.77 -> 33.95 | Helped |
| add local flow perturbations to self-supervised training data (ablation) | PDC-Net-s · KITTI-2015 | AEPE: 7.15 -> 6.66 | Helped |
| add local flow perturbations to self-supervised training data (ablation) | PDC-Net-s · KITTI-2015 | AUSE: 0.256 -> 0.205 | Helped |
| add local flow perturbations to self-supervised training data (ablation) | PDC-Net-s · MegaDepth | PCK-1 (%): 31.53 -> 32.51 | Helped |
| add local flow perturbations to self-supervised training data (ablation) | BaseNet · KITTI-2015 | AEPE: 7.21 -> 7.51 | No clear effect |
| use multi-scale inference | PDC-Net · YFCC100M | mAP @5°: 63.90 -> 65.18 | Helped |
| use probabilistic flow regression instead of the matched GLU-Net-GOCor* baseline | PDC-Net · KITTI-2015 | AEPE: 5.58 -> 5.22 | Helped |
Closest papers by meaning
- Dense Flow from Adaptive Correspondence (NeurIPS 2026 · similarity 0.89)
- UFM: A Simple Path towards Unified Dense Correspondence with Flow (NeurIPS 2025 · similarity 0.88)
- Flow-NeRF: Joint Learning of Geometry, Poses, and Dense Flow within Unified Neural Representations (CVPR 2025 · similarity 0.87)
- Adapting Dense Matching for Homography Estimation with Grid-based Acceleration (CVPR 2025 · similarity 0.85)
- On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity (NeurIPS 2025 · similarity 0.85)
- DifFlow3D: Toward Robust Uncertainty-Aware Scene Flow Estimation with Iterative Diffusion-Based Refinement (CVPR 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.