Research map / Detection and segmentation / Open-set and few-shot detection
PIEs: Pose Invariant Embeddings
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 |
|---|---|---|---|
| Train with pose-invariant view-to-object and object-to-class models instead of multiview classification alone | PI-CNN vs MV-CNN · ModelNet40 | retrieval mAP (%): 47.6 -> 70.0 | Helped |
| Train with pose-invariant view-to-object and object-to-class models instead of multiview classification alone | PI-CNN vs MV-CNN · ObjectPI | classification accuracy (%): 68.1 -> 71.5 | Helped |
| Add pose-invariant view-to-object modeling to multiview triplet-center learning | PI-TC vs MV-TC · ModelNet40 | retrieval mAP (%): 61.4 -> 65.7 | Helped |
| Add pose-invariant view-to-object modeling to multiview triplet-center learning | PI-TC vs MV-TC · ObjectPI | retrieval mAP (%): 62.9 -> 67.4 | Helped |
| Add pose-invariant view-to-object modeling to multiview proxy learning | PI-Proxy vs MV-Proxy · ModelNet40 | classification accuracy (%): 84.7 -> 86.9 | Helped |
| Add pose-invariant view-to-object modeling to multiview proxy learning | PI-Proxy vs MV-Proxy · ObjectPI | classification accuracy (%): 70.7 -> 74.4 | Helped |
| Add pose-invariant view-to-object modeling to multiview methods | PI embeddings vs corresponding MV embeddings · MIRO | classification accuracy (%): 100 -> 100 | No clear effect |
| Use the proposed pose-invariant proxy embedding rather than multiview proxy embedding | PI-Proxy vs MV-Proxy · ObjectPI | retrieval mAP (%): 60.6 -> 63.4 | Helped |
| Use pose-invariant embeddings rather than multiview embeddings | PI embeddings vs corresponding MV embeddings · MIRO | retrieval mAP (%): 92.0; 99.8; 99.8 -> 100 | Mixed |
Closest papers by meaning
- Dual Pose-invariant Embeddings: Learning Category and Object-specific Discriminative Representations for Recognition and Retrieval (CVPR 2024 · similarity 0.93)
- A Comparative Analysis and Study of Multiview CNN Models for Joint Object Categorization and Pose Estimation (ICML 2016 · similarity 0.88)
- ObjectNet: A large-scale bias-controlled dataset for pushing the limits of object recognition models (NeurIPS 2019 · similarity 0.87)
- Thin-Slicing for Pose: Learning to Understand Pose Without Explicit Pose Estimation (CVPR 2016 · similarity 0.87)
- SecondPose: SE(3)-Consistent Dual-Stream Feature Fusion for Category-Level Pose Estimation (CVPR 2024 · similarity 0.86)
- NOPE: Novel Object Pose Estimation from a Single Image (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.