Research map / Label-efficient and robust learning / Knowledge distillation
Distilling Quasi-Conformal Mapping: A Generalizable and Efficient Solution for Wide-Angle Correction
Chengyang Liu, Zixuan Lin, Miaolin Han, Michael K. Ng, Huibin Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| distill quasi-conformal LBS-PGD teacher into QDWC-Net student | QDWC-Net · Tan's test set | inference latency: 26.33 s -> 0.81 s (96.92% reduction) | Improved |
| use annotation-free distilled network instead of Yao et al. trained on 5,133 labelled pairs | QDWC-Net · Tan's test set | FaceAcc: 98.536 -> 98.619 | Improved |
| use QDWC-Net instead of Yao et al. | wide-angle correction · Tan's test set | LineAcc: 70.721 (Yao) vs 70.299 | Hurt |
| use QDWC-Net instead of Yao et al. | wide-angle correction · Tan's test set | BodyAcc: 97.081 -> 97.843 | Improved |
| apply QDWC-Net trained on Tan's data to unseen Zhu's dataset | QDWC-Net · Zhu's dataset | LineAcc: 88.563 -> 88.627 vs Yao | Improved |
| use QDWC-Net instead of Shih et al. on Zhu's dataset | wide-angle correction · Zhu's dataset | FaceAcc: 99.428 (Shih) vs 98.972 | Hurt |
| vary number of Mamba layers, loss terms (Lflow, Limg, Lssim) and output head design (ablation) | QDWC-Net · Tan's test set | correction accuracy | Improved |
Closest papers by meaning
- Lifting the Structural Morphing for Wide-Angle Images Rectification: Unified Content and Boundary Modeling (ICCV 2025 · similarity 0.86)
- QUTCC: Quantile Uncertainty Training and Conformal Calibration for Imaging Inverse Problems (NeurIPS 2026 · similarity 0.83)
- Cylindrical Geodesic Flow Matching for Quasiperiodic Physiological Signal Transformation (NeurIPS 2026 · similarity 0.82)
- Learning from Loss Landscape: Generalizable Mixed-Precision Quantization via Adaptive Sharpness-Aware Gradient Aligning (ICML 2025 · similarity 0.82)
- AdaBM: On-the-Fly Adaptive Bit Mapping for Image Super-Resolution (CVPR 2024 · similarity 0.82)
- Plug-and-Play Tri-Branch Invertible Block for Image Rescaling (AAAI 2025 · similarity 0.81)
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.
