Research map / Image generation / Super-resolution
Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network
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 |
|---|---|---|---|
| Replace fixed input bicubic upscaling with learned sub-pixel convolution after LR-space feature extraction | ESPCN versus SRCNN · Image benchmarks average, scale 3 | PSNR (dB): 27.98 -> 28.11 | Helped |
| Replace fixed input bicubic upscaling with learned sub-pixel convolution after LR-space feature extraction | ESPCN versus SRCNN · Xiph average, scale 3 | PSNR (dB): 33.86 -> 33.92 | Helped |
| Replace fixed input bicubic upscaling with learned sub-pixel convolution after LR-space feature extraction | ESPCN versus SRCNN · Ultra Video Group average, scale 4 | PSNR (dB): 37.52 -> 37.91 | Helped |
| Use tanh instead of ReLU (ablation) | ESPCN trained on ImageNet · Image benchmarks average, scale 3 | PSNR (dB): 28.09 -> 28.11 | Helped |
| Use tanh instead of ReLU (ablation) | ESPCN trained on 91 images · Image benchmarks average, scale 3 | PSNR (dB): 27.76 -> 27.82 | Helped |
| Train on ImageNet instead of 91 images (ablation) | ESPCN with ReLU · Image benchmarks average, scale 3 | PSNR (dB): 27.76 -> 28.09 | Helped |
| Train on ImageNet instead of 91 images (ablation) | SRCNN · Image benchmarks average, scale 3 | PSNR (dB): 27.76 -> 27.83 | Helped |
| Perform most convolutions in LR space and upscale only at the final layer | ESPCN versus SRCNN · 1080p video, scale 3 | seconds per frame: 0.435s -> 0.038s | Helped |
Closest papers by meaning
- Accurate Image Super-Resolution Using Very Deep Convolutional Networks (CVPR 2016 · similarity 0.91)
- Real-Time Video Super-Resolution With Spatio-Temporal Networks and Motion Compensation (CVPR 2017 · similarity 0.91)
- Post-trained Convolution Networks for Single Image Super-resolution (Abstract Reprint) (AAAI 2024 · similarity 0.91)
- Learning a Single Convolutional Super-Resolution Network for Multiple Degradations (CVPR 2018 · similarity 0.91)
- SAVSR: Arbitrary-Scale Video Super-Resolution via a Learned Scale-Adaptive Network (AAAI 2024 · similarity 0.89)
- ResDiff: Combining CNN and Diffusion Model for Image Super-resolution (AAAI 2024 · 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.