Research map / Image generation / Image restoration
One-To-Many Network for Visually Pleasing Compression Artifacts Reduction
Check 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 the one-to-many model with perceptual, naturalness, and JPEG losses instead of the L2 baseline | JPEG artifact reduction network · BSDS500, JPEG Quality 5 | visual detail and human favorability | Helped |
| Add the naturalness loss to the perceptual loss (ablation) | One-to-many network · JPEG artifact reduction | fine details | Helped |
| Add the JPEG loss to the perceptual and naturalness losses (ablation) | One-to-many network · JPEG artifact reduction | contrast consistency | Helped |
| Add the JPEG loss to the perceptual and naturalness losses (ablation) | One-to-many network · JPEG artifact reduction | fine details | Helped |
| Use shift-and-average after deconvolution (ablation) | One-to-many network · JPEG artifact reduction | grid-like artifacts | Helped |
| Remove shift-and-average after deconvolution (ablation) | One-to-many network · JPEG artifact reduction | grid-like artifacts | Hurt |
| Use the perceptual loss alone rather than adding naturalness and JPEG losses (ablation) | One-to-many network · JPEG artifact reduction | fine details | Hurt |
| Use perceptual and naturalness losses without the JPEG loss (ablation) | One-to-many network · JPEG artifact reduction | contrast consistency | Hurt |
| Sample different auxiliary variables to generate multiple candidates | One-to-many network · Set14, JPEG Quality 5 | candidate detail diversity | Helped |
| Use the one-to-many network instead of the compressed JPEG input | One-to-many network · Set14 and BSDS500, JPEG Quality 5 | PSNR | Hurt |
Closest papers by meaning
- Towards image compression with perfect realism at ultra-low bitrates (ICLR 2024 · similarity 0.87)
- Frequency-Biased Synergistic Design for Image Compression and Compensation (CVPR 2025 · similarity 0.87)
- Compress Clean Signal from Noisy Raw Image: A Self-Supervised Approach (ICML 2024 · similarity 0.86)
- Good, Cheap, and Fast: Overfitted Image Compression with Wasserstein Distortion (CVPR 2025 · similarity 0.86)
- Decouple Distortion from Perception: Region Adaptive Diffusion for Extreme-low Bitrate Perception Image Compression (CVPR 2025 · similarity 0.85)
- Controllable Distortion-Perception Tradeoff Through Latent Diffusion for Neural Image Compression (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.