Research map / Image generation / Super-resolution
Downscaled Representation Matters: Improving Image Rescaling with Collaborative Downscaled Images
Bingna Xu, Yong Guo, Luoqian Jiang, Mianjie Yu, Jian Chen
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 collaborative HR and LR example generation at inference | HCFlow · Set5 (4x rescaling) | PSNR (dB): 36.29 -> 36.99 | Improved |
| add collaborative HR and LR example generation at inference | IRN · Set5 (4x rescaling) | PSNR (dB): 36.19 -> 36.63 | Improved |
| add HCD on DIV2K | HCFlow · DIV2K (4x rescaling) | PSNR (dB): 35.32 -> 35.36 | No clear effect |
| add collaborative HR examples to collaborative LR examples (hierarchical) (ablation) | IRN+HCD · Set5 (2x rescaling) | PSNR (dB): 44.45 -> 44.67 | Improved |
| increase collaborative iterations from 1 to 15 (ablation) | IRN+HCD · Set5 (2x rescaling) | PSNR (dB): 44.10 -> 44.67 | Improved |
| increase collaborative iterations from 15 to 20 (ablation) | IRN+HCD · Set5 (2x rescaling) | PSNR (dB): 44.67 -> 44.66 | No clear effect |
| use too low or too high perturbation budget epsilon (ablation) | IRN+HCD · Set5 (2x rescaling) | PSNR (dB) | Hurt |
| optimize LR representation directly with collaborative descent vs adversarial ascent on HCFlow | HCFlow · image rescaling (introduction example) | PSNR gain: +1.29 dB (collaborative) vs -0.93 dB (adversarial) | Improved |
| run HCD only at downscaling stage | IRN+HCD · Set5 (2x rescaling) | upscaling latency (s): 0.03s for all N | No clear effect |
Closest papers by meaning
- Learning Controllable Degradation for Real-World Super-Resolution via Constrained Flows (ICML 2023 · similarity 0.89)
- Task-Aware Image Downscaling (ECCV 2018 · similarity 0.89)
- Deep Generative Model based Rate-Distortion for Image Downscaling Assessment (CVPR 2024 · similarity 0.89)
- Block-Based Multi-Scale Image Rescaling (AAAI 2025 · similarity 0.88)
- Plug-and-Play Tri-Branch Invertible Block for Image Rescaling (AAAI 2025 · similarity 0.87)
- Unsupervised Diffusion-Based Degradation Modeling for Real-World Super-Resolution (AAAI 2025 · similarity 0.85)
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.
