Research map / Image generation / Super-resolution
Implicit Diffusion Models for Continuous 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 IDM instead of SR3 | Face super-resolution · CelebA-HQ 16×16→128×128 | PSNR: 23.04 -> 24.01 | Helped |
| Use IDM instead of GLEAN | Natural image super-resolution · LSUN Cats 16× SR | PSNR: 20.92 -> 21.52 | Helped |
| Use IDM instead of GLEAN | Natural image super-resolution · LSUN Bedrooms 16× SR | PSNR: 19.44 -> 20.33 | Helped |
| Use IDM instead of GLEAN | Natural image super-resolution · LSUN Towers 16× SR | PSNR: 18.41 -> 19.44 | Helped |
| Train IDM with DIV2K and Flicker2K rather than DIV2K only | IDM · DIV2K validation 4× SR | PSNR: 27.10 -> 27.59 | Helped |
| Use IDM instead of LIIF | Continuous face super-resolution · CelebA-HQ 5.3× SR, trained at 8× | LPIPS: 0.1207 -> 0.0526 | Helped |
| Use IDM instead of LIIF | Continuous face super-resolution · CelebA-HQ 5.3× SR, trained at 8× | PSNR: 27.52 -> 23.34 | Hurt |
| Use IDM instead of SR3 | Continuous face super-resolution · CelebA-HQ 10× SR, trained at 8× | PSNR: 20.26 -> 23.46 | Helped |
| Use IDM instead of SR3 | Continuous face super-resolution · CelebA-HQ 10× SR, trained at 8× | LPIPS: 0.2856 -> 0.1171 | Helped |
| Add scale-adaptive modulation using the magnification factor (ablation) | IDM · Face SR at 8× with varied assigned scales | Visual texture detail | Helped |
| Replace the scale-adaptive conditioning network with upsampled LR-image concatenation (ablation) | IDM · LSUN Cats 16× SR | Visual texture detail | Hurt |
| Replace the scale-adaptive conditioning network with encoded LR-feature concatenation (ablation) | IDM · LSUN Cats 16× SR | Visual texture detail | Hurt |
Closest papers by meaning
- Arbitrary-Scale Image Generation and Upsampling using Latent Diffusion Model and Implicit Neural Decoder (CVPR 2024 · similarity 0.92)
- SinSR: Diffusion-Based Image Super-Resolution in a Single Step (CVPR 2024 · similarity 0.91)
- Arbitrary-steps Image Super-resolution via Diffusion Inversion (CVPR 2025 · similarity 0.91)
- Taming Diffusion Prior for Image Super-Resolution with Domain Shift SDEs (NeurIPS 2024 · similarity 0.90)
- Unleashing the Power of One-Step Diffusion based Image Super-Resolution via a Large-Scale Diffusion Discriminator (NeurIPS 2025 · similarity 0.90)
- Effective Diffusion Transformer Architecture for Image Super-Resolution (AAAI 2025 · 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.