Research map / Image generation / Image restoration
From Zero to Detail: A Progressive Spectral Decoupling Paradigm for UHD Image Restoration With New Benchmark
Chen Zhao, Yan Xu, Zhizhou Chen, Enxuan Gu, Kai Zhang, Xiaoming Liu, Jian Jun Yang, Ying Tai
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 |
|---|---|---|---|
| add zero-frequency enhancer (ZFE) stage (ablation) | ERR · UHD-LL | PSNR: 27.01 -> 27.57 | Improved |
| add low-frequency restorer (LFR) stage (ablation) | ERR · UHD-LL | PSNR: 26.81 -> 27.57 | Improved |
| add high-frequency refiner (HFR) stage (ablation) | ERR · UHD-LL | PSNR: 27.04 -> 27.57 | Improved |
| add frequency regularization losses (zero-frequency loss) (ablation) | ERR · UHD-LL | PSNR: 26.73 -> 27.57 | Improved |
| add low-frequency loss (ablation) | ERR · UHD-LL | PSNR: 26.57 -> 27.57 | Improved |
| add high-frequency loss (ablation) | ERR · UHD-LL | PSNR: 26.80 -> 27.57 | Improved |
| add progressive residual between stages (ablation) | ERR · UHD-LL | PSNR: 26.99 -> 27.57 | Improved |
| add global prior projector (ablation) | ERR zero-frequency · UHD-LL | zero-frequency PSNR: 28.15 -> 29.18 | Improved |
| window partition in DCT domain for HFR (add DCT and WP) (ablation) | ERR · UHD-LL | PSNR: 25.78 -> 27.57 | Improved |
| use FW-KAN instead of MLP-24 (ablation) | HFR · UHD-LL | PSNR: 26.97 -> 27.57 | Improved |
| use FW-KAN instead of original KAN (ablation) | HFR · UHD-LL | PSNR: 20.48 -> 27.57 | Improved |
| use three-stage design instead of two-stage V1 (ablation) | ERR · UHD-LL | PSNR: 26.97 -> 27.57 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- RetinexFormer — Retinexformer: One-stage Retinex-based Transformer for Low-light Image Enhancement (ICCV 2023) · 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
- From Zero to Detail: Deconstructing Ultra-High-Definition Image Restoration from Progressive Spectral Perspective (CVPR 2025 · similarity 0.98)
- UHD-processer: Unified UHD Image Restoration with Progressive Frequency Learning and Degradation-aware Prompts (CVPR 2025 · similarity 0.92)
- Latent Harmony: Synergistic Unified UHD Image Restoration via Latent Space Regularization and Controllable Refinement (NeurIPS 2025 · similarity 0.90)
- DreamUHD: Frequency Enhanced Variational Autoencoder for Ultra-High-Definition Image Restoration (AAAI 2025 · similarity 0.89)
- Correlation Matching Transformation Transformers for UHD Image Restoration (AAAI 2024 · similarity 0.89)
- UltraHR-100K: Enhancing UHR Image Synthesis with A Large-Scale High-Quality Dataset (NeurIPS 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.
