Research map / Image generation / Super-resolution
Diffusion Transformer meets Multi-level Wavelet Spectrum for Single Image Super-Resolution
Peng Du, Hui Li, Han Xu, Paul Barom Jeon, Dongwook Lee, Daehyun Ji, Ran Yang, Feng Zhu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| do SR on wavelet spectra instead of pixel domain (ablation) | diffusion transformer · CelebA-HQ 8x face SR | FID: 61.372 -> 60.352 (tokens 1040 -> 704) | Improved |
| add separate LF and HF decoders (ablation) | Freq-DiT · CelebA-HQ 8x face SR | FID: 60.352 -> 58.531 | Improved |
| add LF residual to the dual decoder (ablation) | DTWSR(a) · CelebA-HQ 8x face SR | Cons.: 2.35 -> 0.549 | Improved |
| add LF residual without attention mask (ablation) | DTWSR(a) · CelebA-HQ 8x face SR | FID: 58.531 -> 60.008 | Hurt |
| add tailored attention mask M_high (ablation) | DTWSR(b) · CelebA-HQ 8x face SR | FID: 60.008 -> 56.771 | Improved |
| use pyramid tokenization instead of equal patch size 4 (ablation) | DTWSR · CelebA-HQ 8x face SR | FID: 66.051 -> 56.771 (tokens 1040 -> 704) | Improved |
| allow interactions among multi-scale HF sub-bands (ablation) | DTWSR · CelebA-HQ 8x face SR | FID: 61.670 -> 56.771 | Improved |
| use DTWSR instead of IDM at 16x magnification | face SR · CelebA-HQ 16x | Cons.: 9.57 -> 0.89 | Improved |
| use DTWSR instead of WFEN | face SR 8x · CelebA-HQ 8x | PSNR: 25.53 -> 24.09 | Hurt |
| use DTWSR instead of regression-based HAT | DIV2K 4x · DIV2K val 4x | PSNR: 29.83 -> 28.18 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- ResDiff — ResDiff: Combining CNN and Diffusion Model for Image Super-resolution (AAAI 2024) · better on 3 of 4 shared results
- FlowIE — FlowIE: Efficient Image Enhancement via Rectified Flow (CVPR 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Effective Diffusion Transformer Architecture for Image Super-Resolution (AAAI 2025 · similarity 0.89)
- Accelerated Diffusion via High-Low Frequency Decomposition for Pan-Sharpening (AAAI 2025 · similarity 0.87)
- SinSR: Diffusion-Based Image Super-Resolution in a Single Step (CVPR 2024 · similarity 0.87)
- Training Generative Image Super-Resolution Models by Wavelet-Domain Losses Enables Better Control of Artifacts (CVPR 2024 · similarity 0.87)
- Unleashing the Power of One-Step Diffusion based Image Super-Resolution via a Large-Scale Diffusion Discriminator (NeurIPS 2025 · similarity 0.87)
- One Diffusion Step to Real-World Super-Resolution via Flow Trajectory Distillation (ICML 2025 · 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.
