Research map / Architectures and efficiency / CNN design
CLKN: Cascaded Lucas-Kanade Networks for Image Alignment
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 |
|---|---|---|---|
| use learned convolutional feature pyramids with coarse-to-fine Lucas-Kanade alignment instead of IC-LK | homography estimation · MS-COCO test set | corner error | Helped |
| use learned convolutional feature pyramids with coarse-to-fine Lucas-Kanade alignment instead of SIFT-LK, SIFT+RANSAC, or DHN | homography estimation · MS-COCO test set | corner error | Helped |
| increase the pyramid from 3 to 4 levels (ablation) | CLKN · MS-COCO test set | corner error | Helped |
| increase the pyramid from 4 to 5 levels (ablation) | CLKN · MS-COCO test set | corner error | Hurt |
| use 4 rather than 3 pyramid levels (ablation) | CLKN · MS-COCO test set | number of iterations: 15.6 -> 13.9 | Helped |
| use 4 rather than 5 pyramid levels (ablation) | CLKN · MS-COCO test set | number of iterations: 19.6 -> 13.9 | Helped |
Closest papers by meaning
- Deep Lucas-Kanade Homography for Multimodal Image Alignment (CVPR 2021 · similarity 0.89)
- PRISE: Demystifying Deep Lucas-Kanade With Strongly Star-Convex Constraints for Multimodel Image Alignment (CVPR 2023 · similarity 0.89)
- FastJAM: a Fast Joint Alignment Model for Images (NeurIPS 2025 · similarity 0.87)
- Mnemonic Descent Method: A Recurrent Process Applied for End-To-End Face Alignment (CVPR 2016 · similarity 0.87)
- Inverse Compositional Spatial Transformer Networks (CVPR 2017 · similarity 0.86)
- REPA Works Until It Doesn’t: Early-Stopped, Holistic Alignment Supercharges Diffusion Training (NeurIPS 2025 · similarity 0.84)
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.