Research map / Image generation / Image restoration
Non-Local Recurrent Network for Image Restoration
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 learned non-local feature correlation operations to the recurrent restoration network (ablation) | Parameter-matched RNN · Set12, noise 25 | PSNR: 30.61 -> 30.80 | Helped |
| Share weights across recurrent states (ablation) | NLRN · Set12, noise 25 | PSNR: 30.65 -> 30.80 | Helped |
| Use the non-local module at every recurrent state rather than every other state (ablation) | NLRN · Set12, noise 25 | PSNR: 30.76 -> 30.80 | Helped |
| Use the non-local module at every recurrent state rather than every 3 states (ablation) | NLRN · Set12, noise 25 | PSNR: 30.72 -> 30.80 | Helped |
| Propagate feature correlations between adjacent recurrent states (ablation) | NLRN without correlation propagation · Set12, noise 25 | PSNR: 30.78 -> 30.80 | Helped |
| Use embedded Gaussian similarity rather than Euclidean distance (ablation) | NLRN non-local module · Set12, noise 25 | PSNR: 30.74 -> 30.80 | Helped |
| Use embedded Gaussian similarity rather than dot product (ablation) | NLRN non-local module · Set12, noise 25 | PSNR: 30.68 -> 30.80 | Helped |
| Use embedded Gaussian similarity rather than embedded dot product (ablation) | NLRN non-local module · Set12, noise 25 | PSNR: 30.75 -> 30.80 | Helped |
| Use embedded Gaussian similarity rather than Gaussian similarity (ablation) | NLRN non-local module · Set12, noise 25 | PSNR: 30.69 -> 30.80 | Helped |
| Use embedded Gaussian similarity rather than symmetric embedded Gaussian (ablation) | NLRN non-local module · Set12, noise 25 | PSNR: 30.76 -> 30.80 | Helped |
| Limit correlation computation to a spatial neighborhood instead of using an increasingly broad neighborhood (ablation) | NLRN · Set12, noise 25 | PSNR | Mixed |
Closest papers by meaning
- Image Super-Resolution With Cross-Scale Non-Local Attention and Exhaustive Self-Exemplars Mining (CVPR 2020 · similarity 0.90)
- Non-Local Color Image Denoising With Convolutional Neural Networks (CVPR 2017 · similarity 0.89)
- The Neural Tangent Link Between CNN Denoisers and Non-Local Filters (CVPR 2021 · similarity 0.89)
- Residual Non-local Attention Networks for Image Restoration (ICLR 2019 · similarity 0.89)
- Universal Denoising Networks : A Novel CNN Architecture for Image Denoising (CVPR 2018 · similarity 0.88)
- Omni-Kernel Network for Image Restoration (AAAI 2024 · similarity 0.87)
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.