Research map / Image generation / Autoregressive and latent generation
Low-Level Dataset Distillation for Medical Image Enhancement
Fengzhi Xu, Ziyuan Yang, Mengyu Sun, Joey Tianyi Zhou, Yi Zhang
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 the shared anatomical prior initialization instead of random noise (ablation) | Ours (IPP=5) · CT x4 SR | PSNR: 11.30 -> 34.57 | Improved |
| add the shared anatomical prior initialization instead of random noise (ablation) | Ours (IPP=1) · CT x4 SR | PSNR: 11.54 -> 32.93 | Improved |
| add the pixel-level fidelity preservation step (ablation) | Ours (IPP=5) · CT x4 SR | PSNR: 30.55 -> 34.45 | Improved |
| add the pixel-level fidelity preservation step (ablation) | Ours (IPP=1) · CT x4 SR | PSNR: 16.95 -> 32.73 | Improved |
| increase images per patient from IPP=1 to IPP=5 (ablation) | Ours · CT x4 SR NRI=5 | PSNR: 32.73 -> 34.45 | Improved |
| increase images per patient from IPP=1 to IPP=5 (ablation) | Ours · MRI x2 SR NRI=5 | PSNR: 26.95 -> 28.16 | Improved |
| increase images per patient from IPP=1 to IPP=5 (ablation) | Ours · CT low-dose restoration NRI=5 | PSNR: 28.00 -> 28.38 | Improved |
| increase random real images NRI from 5 to 10 (ablation) | Ours (IPP=5) · CT x4 SR | PSNR: 34.45 -> 34.57 | Improved |
| replace K-Center coreset with distilled data | Ours (IPP=5) · CT x4 SR large-scale (48 patients) NRI=5 | PSNR: 31.75 -> 36.08 | Improved |
| distill with SRCNN and train EDSR instead of K-Center coreset | Ours (IPP=5) · CT x4 SR cross-architecture | PSNR: 34.13 -> 35.48 | Improved |
| distill with SRCNN and train SCTSRN instead of K-Center coreset | Ours (IPP=5) · CT x4 SR cross-architecture | PSNR: 34.47 -> 34.52 | No clear effect |
| replace K-Center coreset with distilled data for restoration | Ours (IPP=5) · CT low-dose restoration NRI=10 | PSNR: 28.11 -> 28.38 | Improved |
Closest papers by meaning
- Rethinking Dataset Distillation for Classification: Do Distilled Sets Outperform Coresets? (NeurIPS 2026 · similarity 0.83)
- What is Dataset Distillation Learning? (ICML 2024 · similarity 0.81)
- What Makes a Good Dataset for Knowledge Distillation? (CVPR 2025 · similarity 0.81)
- Boost Self-Supervised Dataset Distillation via Parameterization, Predefined Augmentation, and Approximation (ICLR 2025 · similarity 0.81)
- DD-Ranking: Rethinking the Evaluation of Dataset Distillation (NeurIPS 2026 · similarity 0.81)
- A Label is Worth A Thousand Images in Dataset Distillation (NeurIPS 2024 · similarity 0.81)
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.
