Research map / Label-efficient and robust learning / Semi-supervised learning
Semi-supervised Deep Large-Baseline Homography Estimation with Progressive Equivalence Constraint
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 |
|---|---|---|---|
| Apply progressive estimation with two inserted intermediate images and the semi-supervised homography identity loss | Proposed homography estimator · Large-baseline dataset (Avg-L) | PME: 10.30 -> 5.16 | Helped |
| Apply the proposed estimator | Proposed homography estimator · Small-baseline dataset (Avg-S) | PME: 0.50 -> 0.44 | Helped |
| Use one rather than two inserted intermediate images (ablation) | Progressive estimation strategy · Large-baseline dataset (Avg-L) | PME: 5.16 -> 5.50 | Hurt |
| Use three rather than two inserted intermediate images (ablation) | Progressive estimation strategy · Large-baseline dataset (Avg-L) | PME: 5.16 -> 6.42 | Hurt |
| Replace the homography identity loss with photometric loss (ablation) | Proposed homography estimator · Large-baseline dataset (Avg-L) | PME: 5.16 -> 53.53 | Hurt |
| Replace the homography identity loss with ablation-based photometric loss (ablation) | Proposed homography estimator · Large-baseline dataset (Avg-L) | PME: 5.16 -> 12.23 | Hurt |
| Remove the unsupervised objective from the semi-supervised loss (ablation) | Homography identity loss · Large-baseline dataset (Avg-L) | PME: 5.16 -> 7.50 | Hurt |
| Remove the supervised objective from the semi-supervised loss (ablation) | Homography identity loss · Large-baseline dataset (Avg-L) | PME: 5.16 -> 6.16 | Hurt |
| Use only the unsupervised objective instead of only the supervised objective (ablation) | Homography identity loss · Large-baseline dataset (Avg-L) | PME: 7.50 -> 6.16 | Helped |
| Use corner offsets instead of homography flow (ablation) | Homography flow training representation · Large-baseline dataset (Avg-L) | PME: 5.16 -> 11.51 | Hurt |
| Use 8 bases flow instead of homography flow (ablation) | Homography flow training representation · Large-baseline dataset (Avg-L) | PME: 5.16 -> 12.09 | Hurt |
Closest papers by meaning
- MCNet: Rethinking the Core Ingredients for Accurate and Efficient Homography Estimation (CVPR 2024 · similarity 0.90)
- Adapting Dense Matching for Homography Estimation with Grid-based Acceleration (CVPR 2025 · similarity 0.90)
- Noisy One-point Homographies are Surprisingly Good (CVPR 2024 · similarity 0.87)
- SSHNet: Unsupervised Cross-modal Homography Estimation via Problem Reformulation and Split Optimization (CVPR 2025 · similarity 0.87)
- Mask-Homo: Pseudo Plane Mask-Guided Unsupervised Multi-Homography Estimation (AAAI 2024 · similarity 0.87)
- Unsupervised Homography Estimation on Multimodal Image Pair via Alternating Optimization (NeurIPS 2024 · similarity 0.87)
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.