Research map / Label-efficient and robust learning / Domain adaptation and generalization
Multi-Domain Multi-Definition Landmark Localization for Small Datasets
David Ferman, Gaurav Bharaj
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 300W concurrent training (ablation) | MDMD · AnimWeb | NME_ic: 6.88 -> 6.55 | Improved |
| add 300W concurrent training (ablation) | MDMD · ArtFace | NME_ic: 4.46 -> 3.75 | Improved |
| add 300W concurrent training (ablation) | MDMD · CariFace | NME_ic: 7.81 -> 5.85 | Improved |
| add 300W concurrent training (ablation) | MDMD · PARE | NME_ic: 9.12 -> 8.59 | Improved |
| add LaPa concurrent training | MDMD · COFW | NME_ip: 4.82 -> 4.65 | Improved |
| add LaPa concurrent training | MDMD · WFLW | NME_ic: 4.06 -> 3.97 | Improved |
| use Laplacian log-likelihood instead of Euclidean loss (ablation) | MDMD Single · COFW | NME_ip: 4.90 -> 4.82 | Improved |
| use FLSG tokens instead of individual landmark tokens (ablation) | MDMD Single · COFW | NME_ip: 4.73 -> 4.82 | Hurt |
| use 12 FLSG groups instead of 5 (ablation) | MDMD · WFLW | NME_ic: 4.12 -> 4.06 | Improved |
| pretrain ViT instead of training from scratch (ablation) | MDMD · COFW | NME_ip: 13.97 -> 4.82 | Improved |
| use ViT instead of ResNet-50 backbone (ablation) | MDMD · COFW | NME_ip: 5.10 -> 4.82 | Improved |
| use MDMD joint training instead of sequential transfer learning | MDMD · WFLW with LaPa | NME_ic: 4.00 -> 3.97 | No clear effect |
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.
- ADNet — ADNet: Leveraging Error-Bias Towards Normal Direction in Face Alignment (ICCV 2021) · better on 6 of 7 shared results
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
- Improving Landmark Localization With Semi-Supervised Learning (CVPR 2018 · similarity 0.90)
- Joint Super-Resolution and Alignment of Tiny Faces (AAAI 2020 · similarity 0.88)
- FaceLift: Semi-supervised 3D Facial Landmark Localization (CVPR 2024 · similarity 0.87)
- Laplace Landmark Localization (ICCV 2019 · similarity 0.87)
- AnchorFace: An Anchor-based Facial Landmark Detector Across Large Poses (AAAI 2021 · similarity 0.87)
- LAFS: Landmark-based Facial Self-supervised Learning for Face Recognition (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.
