Research map / Detection and segmentation / CNN features
Neural Outlier Rejection for Self-Supervised Keypoint Learning
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 |
|---|---|---|---|
| Enable cross-border keypoint detection (ablation) | KeyPointNet · HPatches, 240x320 | Repeatability: 0.633 -> 0.689 | Helped |
| Enable cross-border keypoint detection (ablation) | KeyPointNet · HPatches, 240x320 | M.Score: 0.491 -> 0.537 | Helped |
| Add descriptor upsampling (ablation) | KeyPointNet with cross-border detection · HPatches, 240x320 | Cor-1: 0.491 -> 0.579 | Helped |
| Add descriptor upsampling (ablation) | KeyPointNet with cross-border detection · HPatches, 240x320 | Cor-3: 0.805 -> 0.866 | Helped |
| Train descriptors using L_IO without L_desc (ablation) | KeyPointNet with descriptor upsampling · HPatches, 240x320 | Cor-1: 0.579 -> 0.602 | Helped |
| Train descriptors using L_IO without L_desc (ablation) | KeyPointNet with descriptor upsampling · HPatches, 240x320 | Cor-3: 0.866 -> 0.836 | Hurt |
| Add triplet descriptor loss L_desc (ablation) | V3 - IO-Net · HPatches, 240x320 | Cor-3: 0.836 -> 0.867 | Helped |
| Add triplet descriptor loss L_desc (ablation) | V3 - IO-Net · HPatches, 240x320 | Cor-1: 0.602 -> 0.591 | Hurt |
| Use the proposed self-supervised keypoint method instead of UnsuperPoint | HPatches · HPatches, 240x320 | Repeatability: 0.645 -> 0.686 | Helped |
| Use the proposed self-supervised keypoint method instead of UnsuperPoint | HPatches · HPatches, 240x320, 300 points | M.Score: 0.424 -> 0.544 | Helped |
Closest papers by meaning
- Self-Supervised Equivariant Learning for Oriented Keypoint Detection (CVPR 2022 · similarity 0.87)
- Key-Grid: Unsupervised 3D Keypoints Detection using Grid Heatmap Features (NeurIPS 2024 · similarity 0.85)
- KeyPose: Category-Level 6D Object Pose Estimation with Self-Adaptive Keypoints (AAAI 2025 · similarity 0.85)
- Back to 3D: Few-Shot 3D Keypoint Detection with Back-Projected 2D Features (CVPR 2024 · similarity 0.85)
- Meta-Point Learning and Refining for Category-Agnostic Pose Estimation (CVPR 2024 · similarity 0.84)
- 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.