Research map / Detection and segmentation / Open-set and few-shot detection
Self-Supervised Learning of Geometrically Stable Features Through Probabilistic Introspection
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 |
|---|---|---|---|
| Add probabilistic confidence prediction (ablation) | Ours · PF-Pascal test | PCK@0.1: 60.6 -> 66.5 | Helped |
| Use probabilistic introspection rather than the contrastive loss (ablation) | Ours · PF-Pascal semantic matching | region matching performance | Helped |
| Train with all 20 PASCAL categories, including non-rigid classes (ablation) | Ours · PF-Pascal proposal matching | proposal matching performance | No clear effect |
| Train with all 20 PASCAL categories, including non-rigid classes (ablation) | Ours · Pascal3D few-shot keypoint detection | area under PCK-over-alpha curve | Hurt |
| Use ground-truth boxes instead of weakly supervised detections (ablation) | Ours · PF-Pascal semantic matching and Pascal3D keypoint detection | task performance | No clear effect |
| Remove bounding-box annotations (ablation) | Ours · PF-Pascal and Pascal3D | task performance | Mixed |
Closest papers by meaning
- Unsupervised Learning of Object Landmarks via Self-Training Correspondence (NeurIPS 2020 · similarity 0.90)
- Self-Supervised Learning of Object Parts for Semantic Segmentation (CVPR 2022 · similarity 0.88)
- AnchorNet: A Weakly Supervised Network to Learn Geometry-Sensitive Features for Semantic Matching (CVPR 2017 · similarity 0.88)
- Reinforced Feature Points: Optimizing Feature Detection and Description for a High-Level Task (CVPR 2020 · similarity 0.88)
- Object landmark discovery through unsupervised adaptation (NeurIPS 2019 · similarity 0.88)
- Pose-Guided Self-Training with Two-Stage Clustering for Unsupervised Landmark Discovery (CVPR 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.