Research map / Image generation / Super-resolution
Amortised MAP Inference for Image Super-resolution
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 |
|---|---|---|---|
| Use affine projected GAN instead of the soft-constrained GAN (ablation) | GAN-based super-resolution · CelebA | PSNR | Helped |
| Use affine projection instead of a soft consistency constraint (ablation) | GAN-based super-resolution · CelebA | input-output consistency | Helped |
| Add affine projection to the CNN (ablation) | MSE-trained CNN · CelebA | MSE | Helped |
| Add affine projection to the CNN (ablation) | MSE-trained CNN · CelebA | SSIM | Helped |
| Initialize the affine projection with the correct Moore-Penrose inverse (ablation) | Affine projected CNN · CelebA | MSE | Helped |
| Train the correctly initialized A+ instead of keeping it fixed (ablation) | Affine projected CNN · CelebA | MSE | No clear effect |
| Randomly initialize the final A+ deconvolution instead of using the correct inverse initialization (ablation) | Affine projected CNN · CelebA | MSE | Hurt |
| Use instance noise during GAN training (ablation) | AffGAN · ImageNet | training stability | Helped |
| Use GAN training rather than MSE training (ablation) | Super-resolution model · CelebA | PSNR | Hurt |
| Use GAN training rather than MSE training (ablation) | Super-resolution model · CelebA | perceptual sharpness | Helped |
| Use denoiser-guided or density-guided optimization instead of AffGAN (ablation) | Affine projected super-resolution · Grass textures | visual sharpness | Hurt |
| Use the affine projection instead of the soft constraint (ablation) | Super-resolution model · 2D Swiss-roll | cross-entropy | Helped |
Closest papers by meaning
- SinSR: Diffusion-Based Image Super-Resolution in a Single Step (CVPR 2024 · similarity 0.87)
- Noise-Started One-Step Image Super-Resolution via LR-Conditioned SplitMeanFlow and GAN Refinement (NeurIPS 2026 · similarity 0.87)
- AdaBM: On-the-Fly Adaptive Bit Mapping for Image Super-Resolution (CVPR 2024 · similarity 0.86)
- GSDD: Generative Space Dataset Distillation for Image Super-resolution (AAAI 2024 · similarity 0.86)
- ResDiff: Combining CNN and Diffusion Model for Image Super-resolution (AAAI 2024 · similarity 0.86)
- SeD: Semantic-Aware Discriminator for Image Super-Resolution (CVPR 2024 · similarity 0.86)
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.