Research map / Image generation / Super-resolution
Learning with Privileged Information for Efficient Image Super-Resolution
Wonkyung Lee, Junghyup Lee, Dohyung Kim, Bumsub Ham
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 |
|---|---|---|---|
| add feature distillation from a teacher fed with HR images (plain teacher) (ablation) | FSRCNN student · Set5 x2 | PSNR (dB): 37.15 -> 37.19 | Improved |
| add hourglass architecture to the teacher (ablation) | FSRCNN student · Set5 x2 | PSNR (dB): 37.19 -> 37.22 | Improved |
| add weight transfer from teacher decoder to student (ablation) | FSRCNN student · Set5 x2 | PSNR (dB): 37.22 -> 37.23 | Improved |
| add imitation loss to teacher (ablation) | FSRCNN student · Set5 x2 | PSNR (dB): 37.23 -> 37.27 | Improved |
| use VID Gaussian distillation loss instead of MAE (ablation) | FSRCNN student · Set5 x2 | PSNR (dB): 37.27 -> 37.31 | Improved |
| use VID Laplace instead of Gaussian distillation loss (ablation) | FSRCNN student · Set5 x2 | PSNR (dB): 37.31 -> 37.33 | Improved |
| full framework vs undistilled baseline | FSRCNN · Set5 x2 | PSNR (dB): 37.15 -> 37.33 | Improved |
| hourglass limits teacher quality, narrowing teacher-student gap (ablation) | teacher · Set5 x2 | teacher PSNR (dB): 57.60 -> 37.70 | Hurt |
| apply the framework to FSRCNN | FSRCNN* · Urban100 x2 | PSNR (dB): 30.05 -> 30.24 | Improved |
| apply the framework to FSRCNN | FSRCNN* · Urban100 x4 | PSNR (dB): 24.76 -> 24.82 | Improved |
| apply the framework to FSRCNN | FSRCNN* · Set5 x4 | PSNR (dB): 30.89 -> 30.95 | Improved |
| apply the framework to IDN | IDN · Set5 x4 | PSNR (dB) | No clear effect |
Closest papers by meaning
- Image Super-Resolution Using Deep Convolutional Networks (TPAMI 2016 · similarity 0.92)
- Learning a Single Convolutional Super-Resolution Network for Multiple Degradations (CVPR 2018 · similarity 0.92)
- Accelerating the Super-Resolution Convolutional Neural Network (ECCV 2016 · similarity 0.91)
- Accurate Image Super-Resolution Using Very Deep Convolutional Networks (CVPR 2016 · similarity 0.91)
- Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network (CVPR 2016 · similarity 0.91)
- Deep Learning for Image Super-Resolution: A Survey (TPAMI 2020 · similarity 0.90)
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.
