Research map / Label-efficient and robust learning / Knowledge distillation
Self-Decoupling and Ensemble Distillation for Efficient Segmentation
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 DPED to segmentation training (ablation) | DeepLabV3-EfficientNet-B0 with SDES · CamVid | mIoU: 67.56 -> 68.13 | Helped |
| Add DFED to segmentation training (ablation) | DeepLabV3-EfficientNet-B0 with SDES · CamVid | mIoU: 67.56 -> 68.37 | Helped |
| Add both DPED and DFED (ablation) | DeepLabV3-EfficientNet-B0 with SDES · CamVid | mIoU: 67.56 -> 68.74 | Helped |
| Use a random category split instead of the more-to-less subset split (ablation) | SDES · CamVid | mIoU: 68.74 -> 68.19 | Hurt |
| Vary the DPED loss weight α (ablation) | SDES · CamVid | mIoU | Mixed |
| Vary the DFED loss weight β (ablation) | SDES · CamVid | mIoU | Mixed |
| Vary the feature selection ratio γ (ablation) | SDES · CamVid | mIoU | Mixed |
| Vary the number of expert decoders N (ablation) | SDES · CamVid | mIoU | Mixed |
Closest papers by meaning
- Refine Myself by Teaching Myself: Feature Refinement via Self-Knowledge Distillation (CVPR 2021 · similarity 0.91)
- SDPGO: Efficient Self-Distillation Training Meets Proximal Gradient Optimization (NeurIPS 2025 · similarity 0.89)
- Distilling Knowledge from Heterogeneous Architectures for Semantic Segmentation (AAAI 2025 · similarity 0.88)
- Scaled Decoupled Distillation (CVPR 2024 · similarity 0.87)
- Multi-label Self Knowledge Distillation (AAAI 2025 · similarity 0.87)
- How to Trade Off the Quantity and Capacity of Teacher Ensemble: Learning Categorical Distribution to Stochastically Employ a Teacher for Distillation (AAAI 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.
