Research map / Label-efficient and robust learning / Knowledge distillation
Knowledge Distillation for 6D Pose Estimation by Aligning Distributions of Local Predictions
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 |
|---|---|---|---|
| Align teacher and student local prediction distributions using optimal transport | WDRNet+ with DarkNet-tiny · LINEMOD | ADD-0.1d: 88.7 -> 89.9 | Helped |
| Align teacher and student local prediction distributions using optimal transport | WDRNet+ with DarkNet-tiny-H · LINEMOD | ADD-0.1d: 81.9 -> 84.2 | Helped |
| Use naive prediction-to-prediction distillation | WDRNet+ with DarkNet-tiny-H · LINEMOD | ADD-0.1d: 81.9 -> 81.6 | Hurt |
| Combine local prediction distribution alignment with FKD | WDRNet+ with DarkNet-tiny-H · LINEMOD | ADD-0.1d: 81.9 -> 84.8 | Helped |
| Align local prediction distributions | WDRNet+ · Occluded-LINEMOD | ADD-0.1d: 40.5 -> 43.2 | Helped |
| Combine local prediction distribution alignment with FKD | WDRNet+ · Occluded-LINEMOD | ADD-0.1d: 40.5 -> 44.5 | Helped |
| Align local prediction distributions | WDRNet+ · YCB-V | ADD-0.1d: 16.1 -> 18.7 | Helped |
| Align dense local prediction distributions | ZebraPose · Occluded-LINEMOD | ADD-0.1d: 71.4 -> 73.3 | Helped |
| Add segmentation scores to optimal-transport distribution alignment (ablation) | WDRNet+ with DarkNet-tiny · LINEMOD | ADD-0.1d: 89.1 -> 89.9 | Helped |
| Add segmentation scores to optimal-transport distribution alignment (ablation) | WDRNet+ with DarkNet-tiny-H · LINEMOD | ADD-0.1d: 83.1 -> 84.2 | Helped |
| Apply distribution alignment without detection preprocessing (ablation) | WDRNet · LINEMOD | ADD-0.1d: 66.8 -> 68.9 | Helped |
| Apply distribution alignment with a PnP network (ablation) | WDRNet + PnPNet · LINEMOD | ADD-0.1d: 62.0 -> 69.8 | Helped |
Closest papers by meaning
- Single-Stage 6D Object Pose Estimation (CVPR 2020 · similarity 0.90)
- FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (CVPR 2024 · similarity 0.89)
- Segmentation-Driven 6D Object Pose Estimation (CVPR 2019 · similarity 0.89)
- PoseAgent: Budget-Constrained 6D Object Pose Estimation via Reinforcement Learning (CVPR 2017 · similarity 0.88)
- Keypoint-Graph-Driven Learning Framework for Object Pose Estimation (CVPR 2021 · similarity 0.88)
- OVE6D: Object Viewpoint Encoding for Depth-Based 6D Object Pose Estimation (CVPR 2022 · similarity 0.88)
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.