Research map / 3D vision / Depth and stereo
Robust Homography Estimation via Dual Principal Component Pursuit
Tianjiao Ding, Yunchen Yang, Zhihui Zhu, Daniel P. Robinson, Rene Vidal, Laurent Kneip, Manolis C. Tsakiris
Check 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 DPCP(1) with homographic embeddings instead of USAC with local optimization | Robust homography estimation · MS-COCO warped images, mean | time: 16.6 -> 1.13 | Improved |
| Use DPCP(1) with homographic embeddings instead of USAC with local optimization | Robust homography estimation · MS-COCO warped images, mean | corner error: 1.18 -> 2.53 | Hurt |
| Use epipolar embeddings and recover the homography from a 3-dimensional nullspace instead of homographic embeddings (ablation) | DPCP homography estimation · MS-COCO warped images, mean | corner error: 2.53 -> 5.08 | Hurt |
| Use epipolar embeddings and recover the homography from a 3-dimensional nullspace instead of homographic embeddings (ablation) | DPCP homography estimation · MS-COCO warped images, median | corner error: 0.81 -> 0.96 | Hurt |
| Disable progressive sampling in USAC while retaining local optimization (ablation) | USAC homography estimation · MS-COCO warped images, mean | corner error: 6.75 -> 1.18 | Improved |
| Use DPCP(1) instead of USAC with local optimization | Two-view homography estimation · KITTI-8 | time: 13.1 -> 0.64 | Improved |
| Use DPCP(1) instead of USAC with local optimization | Two-view homography estimation · KITTI-8 | rotation error: 0.42 -> 0.52 | Hurt |
| Estimate jointly from three views using DPCP(1) instead of using two-view DPCP(1) (ablation) | DPCP homography estimation · KITTI-8 | rotation error: 0.52 -> 0.35 | Improved |
| Use DPCP(1) instead of USAC with local optimization | Two-view homography estimation · TUM-near | time: 30.6 -> 0.84 | Improved |
| Use DPCP(1) instead of USAC with local optimization | Two-view homography estimation · TUM-near | rotation error: 0.42 -> 0.46 | Hurt |
| Use DPCP(6) with trifocal embeddings instead of DPCP(1) with homographic embeddings (ablation) | Three-view homography estimation · TUM-hemisphere | rotation error: 2.46 -> 1.32 | Hurt |
| Use DPCP(6) with trifocal embeddings instead of DPCP(1) with homographic embeddings (ablation) | Three-view homography estimation · TUM-hemisphere | time: 4.12 -> 2.70 | Improved |
Closest papers by meaning
- Noisy One-point Homographies are Surprisingly Good (CVPR 2024 · similarity 0.90)
- Adapting Dense Matching for Homography Estimation with Grid-based Acceleration (CVPR 2025 · similarity 0.87)
- Unsupervised Homography Estimation on Multimodal Image Pair via Alternating Optimization (NeurIPS 2024 · similarity 0.87)
- Mask-Homo: Pseudo Plane Mask-Guided Unsupervised Multi-Homography Estimation (AAAI 2024 · similarity 0.86)
- MCNet: Rethinking the Core Ingredients for Accurate and Efficient Homography Estimation (CVPR 2024 · similarity 0.86)
- PARSAC: Accelerating Robust Multi-Model Fitting with Parallel Sample Consensus (AAAI 2024 · similarity 0.86)
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.
