Research map / Label-efficient and robust learning / Semi-supervised learning
Semi-supervised Keypoint Localization
Olga Moskvyak, Frederic Maire, Feras Dayoub, Mahsa Baktashmotlagh
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 semantic consistency to the TE + TI losses (ablation) | HRNet-32 semi-supervised keypoint localization · LSP (5% labeled) | PCK@0.1: 46.76 -> 66.32 | Improved |
| add semantic consistency to the supervised baseline (ablation) | HRNet-32 · LSP (5% labeled) | PCK@0.1: 39.16 -> 66.32 | Improved |
| add transformation equivariance and transformation invariance to SC (ablation) | HRNet-32 semi-supervised keypoint localization · LSP (5% labeled) | PCK@0.1: 64.74 -> 66.32 | Improved |
| add transformation equivariance to TI + SC (ablation) | HRNet-32 semi-supervised keypoint localization · LSP (5% labeled) | PCK@0.1: 65.23 -> 66.32 | Improved |
| add transformation invariance to TE + SC (ablation) | HRNet-32 semi-supervised keypoint localization · LSP (5% labeled) | PCK@0.1: 65.78 -> 66.32 | Improved |
| add semantic consistency and transformation equivariance to TI (ablation) | HRNet-32 semi-supervised keypoint localization · LSP (5% labeled) | PCK@0.1: 43.62 -> 66.32 | Improved |
| use ground-truth heatmaps in unsupervised losses (ablation) | HRNet-32 semi-supervised keypoint localization · LSP (5% labeled) | PCK@0.1: 64.75 -> 66.32 | Improved |
| increase the amount of unlabeled training data (ablation) | HRNet-32 semi-supervised keypoint localization · CUB-200-2011 (5% labeled) | PCK@0.1: 87.01 -> 91.34 | Improved |
| increase the amount of unlabeled training data (ablation) | HRNet-32 semi-supervised keypoint localization · ATRW (5% labeled) | PCK@0.1: 72.04 -> 93.02 | Improved |
| replace transformation equivariance with inverse transformation loss (ablation) | HRNet-32 semi-supervised keypoint localization · ATRW and CUB-200-2011 (20% labeled) | PCK@0.1: decreased by 1% | Hurt |
| use affine rather than perspective transformations (ablation) | HRNet-32 semi-supervised keypoint localization · keypoint localization benchmarks | PCK | Hurt |
| collect additional intermediate features (ablation) | HRNet-32 semi-supervised keypoint localization · keypoint localization benchmarks | PCK | No clear effect |
Closest papers by meaning
- Few-Shot Geometry-Aware Keypoint Localization (CVPR 2023 · similarity 0.89)
- Few-Shot Keypoint Detection With Uncertainty Learning for Unseen Species (CVPR 2022 · similarity 0.89)
- Detect Any Keypoints: An Efficient Light-Weight Few-Shot Keypoint Detector (AAAI 2024 · similarity 0.86)
- Key-Grid: Unsupervised 3D Keypoints Detection using Grid Heatmap Features (NeurIPS 2024 · similarity 0.85)
- Learning to Produce Semi-dense Correspondences for Visual Localization (CVPR 2024 · similarity 0.85)
- Recurrent Feature Mining and Keypoint Mixup Padding for Category-Agnostic Pose Estimation (CVPR 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.
