Research map / Image generation / Image restoration
The Perception-Distortion Tradeoff
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 |
|---|---|---|---|
| Require lower average distortion | Any image-restoration estimator · Image restoration | perceptual-quality index | Hurt |
| Use VGG2,2 feature distance instead of pixel-based distortion measures | Super-resolution algorithms · BSD100, 4× super-resolution | perception-distortion tradeoff | Helped |
| Add an adversarial loss and vary its weight alongside the distortion loss | Wasserstein GAN denoiser · MNIST denoising, σ = 3 | perception-distortion tradeoff | Helped |
| Use the denoising WGAN estimator instead of MAP estimation | Denoising estimator · MNIST denoising, σ = 3 | perceptual quality | Helped |
| Use the denoising WGAN estimator instead of randomly drawing images from the dataset | Denoising estimator · MNIST denoising, σ = 3 | distortion | Helped |
| Optimize VGG2,2 feature distance rather than MSE (ablation) | SRResNet · BSD100, 4× super-resolution | image sharpness | Helped |
| Optimize VGG2,2 feature distance rather than MSE (ablation) | SRResNet · BSD100, 4× super-resolution | periodic artifacts | Hurt |
| Add an adversarial loss to the distortion loss (ablation) | SRGAN · BSD100, 4× super-resolution | perceptual quality | Helped |
Closest papers by meaning
- Traversing Distortion-Perception Tradeoff using a Single Score-Based Generative Model (CVPR 2025 · similarity 0.89)
- Posterior-Mean Rectified Flow: Towards Minimum MSE Photo-Realistic Image Restoration (ICLR 2025 · similarity 0.88)
- Augmenting Perceptual Super-Resolution via Image Quality Predictors (CVPR 2025 · similarity 0.87)
- Looks Too Good To Be True: An Information-Theoretic Analysis of Hallucinations in Generative Restoration Models (NeurIPS 2024 · similarity 0.87)
- UARE: A Unified Vision-Language Model for Image Quality Assessment, Restoration, and Enhancement (CVPR 2026 · similarity 0.87)
- Evaluation of Visual Processing Should Be Human-Centered, Not Metric-Centered (NeurIPS 2026 · 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.