Research map / 3D vision / Depth and stereo
RANSAC-Flow: Generic Two-stage Image Alignment
Xi Shen, François Darmon, Alexei A. Efros, Mathieu Aubry
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 multiple homographies (ablation) | RANSAC-Flow (ImageNet features) · KITTI 2015 train (non-occluded) | AEE: 6.74 -> 3.87 | Improved |
| add multiple homographies (ablation) | RANSAC-Flow (MoCo features) · KITTI 2015 train (non-occluded) | AEE: 7.04 -> 4.15 | Improved |
| add multiple homographies (ablation) | RANSAC-Flow (ImageNet features) · MegaDepth sparse correspondences (3 px) | accuracy: 77.84 -> 83.34 | Improved |
| add multiple homographies (ablation) | RANSAC-Flow (MoCo features) · YFCC100M two-view geometry (mAP@5deg) | mAP: 61.10 -> 64.88 | Improved |
| add multiple homographies (ablation) | RANSAC-Flow (ImageNet features) · YFCC100M two-view geometry (mAP@5deg) | mAP: 59.90 -> 62.45 | Improved |
| add multiple homographies (ablation) | RANSAC-Flow (MoCo features) · Aachen night-time localization (0.5m, 2deg) | localization accuracy: 42.9 -> 44.9 | Improved |
| add fine-tuning on the target set (ablation) | RANSAC-Flow (MoCo features) · KITTI 2015 train (non-occluded) | AEE: 4.67 -> 4.15 | Improved |
| add fine-tuning on the target set (ablation) | RANSAC-Flow (ImageNet features) · MegaDepth sparse correspondences (3 px) | accuracy: 83.31 -> 83.34 | No clear effect |
| add fine-tuning on the target set (ablation) | RANSAC-Flow (MoCo features) · Aachen night-time localization (0.5m, 2deg) | localization accuracy: 41.8 -> 44.9 | Improved |
| add fine-tuning on the target set (ablation) | RANSAC-Flow (MoCo features) · YFCC100M two-view geometry (mAP@5deg) | mAP: 63.48 -> 64.88 | Improved |
| add fine flow network to coarse homography | RANSAC-Flow (ImageNet features) · Hpatches viewpoint 5 | AEE: 10.07 -> 5.12 | Improved |
| use MoCo instead of ImageNet features | RANSAC-Flow · YFCC100M two-view geometry (mAP@5deg) | mAP: 62.45 -> 64.88 | Improved |
Later papers that compared with this method
2 later accepted papers report RANSAC-Flow as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Co-Attention for Conditioned Image Matching (CVPR 2021 · better on 0 of 2 shared results)
- Warp Consistency for Unsupervised Learning of Dense Correspondences (ICCV 2021 · better on 2 of 2 shared results)
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.
- OANet — OANet: Learning Two-View Correspondences and Geometry Using Order-Aware Network (TPAMI 2020) · better on 1 of 1 shared result
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
- Dense Cross-Scale Image Alignment with Fully Spatial Correlation and Just Noticeable Difference Guidance (AAAI 2026 · similarity 0.89)
- FastJAM: a Fast Joint Alignment Model for Images (NeurIPS 2025 · similarity 0.86)
- Jointly Aligning Millions of Images With Deep Penalised Reconstruction Congealing (ICCV 2019 · similarity 0.86)
- Video Stitching from Multiple Moving Cameras (NeurIPS 2026 · similarity 0.86)
- Content-Aware Unsupervised Deep Homography Estimation (ECCV 2020 · similarity 0.85)
- Hierarchical Feature Alignment Network for Unsupervised Video Object Segmentation (ECCV 2022 · 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.
