Research map / Continual learning / Online and replay-based
Expandable, Compressible, Mineable: Open-World Thermal Infrared Image Restoration
Pu Li, Huafeng Li, Yafei Zhang, Wen Wang, Neng Dong, Jie Wen
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 |
|---|---|---|---|
| use group-isolated expansion, SEP and SKMM together instead of a single shared backbone (ablation) | ECMRNet · HM-TIR average | PSNR: 26.21 -> 30.36 | Improved |
| add Structural Entropy Pruning (ablation) | ECMRNet · HM-TIR average | PSNR / Params: Params 27.51M -> 19.40M, PSNR 30.51 -> 30.36 | Improved |
| add SKMM knowledge mining for compound degradations (ablation) | ECMRNet · HM-TIR compound degradations | PSNR: 25.44 -> 26.36 | Improved |
| add SKMM (ablation) | ECMRNet · HM-TIR compound degradations | FLOPs: 76.12G -> 297.97G | Hurt |
| raise SEP frequency threshold rho from 0.10 to 0.20 (ablation) | SEP · HM-TIR contrast | PSNR: 37.46 -> 36.34 | Hurt |
| lower SEP frequency threshold rho from 0.10 to 0.05 (ablation) | SEP · HM-TIR contrast | Params / PSNR: PSNR 37.46 -> 37.54, Params 2.95M -> 3.59M | Mixed |
| increase SEP sample pool size N from 100 to 400 (ablation) | SEP · HM-TIR contrast | PSNR: 37.46 -> 37.47 | No clear effect |
| use SEP instead of magnitude pruning or sparsity-regularized pruning (ablation) | channel-group pruning · HM-TIR (no fine-tuning, avg) | PSNR: MP 18.90, SRP 15.75 -> SEP 25.71 | Improved |
| apply SEP to longer task sequence (dead pixels task) (ablation) | ECMRNet · HM-TIR dead pixels | PSNR / Params: PSNR 45.34 -> 45.10, Params 4.07M -> 2.37M | Improved |
| use ECMRNet instead of all-in-one restorers and continual-learning baselines | TIR restoration · HM-TIR average | PSNR: 28.77 -> 30.36 (vs best baseline MoCE-IR) | Improved |
| use regularization-based continual learning (LwF, EWC, MAS, SI) on NAFNet | NAFNet · HM-TIR average | PSNR: NAFNet 26.39 -> LwF 22.12, EWC 21.48, MAS 22.34, SI 22.03 | Hurt |
Closest papers by meaning
- Toward Real-world Infrared Image Super-Resolution: A Unified Autoregressive Framework and Benchmark Dataset (CVPR 2026 · similarity 0.88)
- Discovering Unseen Degradations to Adapt Open-World Image Restoration (NeurIPS 2026 · similarity 0.88)
- HiDRA: Hierarchical Degradation Representation and Adaptation with Generative Priors for Enhancing Infrared Vision (CVPR 2026 · similarity 0.88)
- Enhancing Infrared Vision: Progressive Prompt Fusion Network and Benchmark (NeurIPS 2025 · similarity 0.86)
- TherA: Thermal-Aware Visual-Language Prompting for Controllable RGB-to-Thermal Infrared Translation (CVPR 2026 · similarity 0.86)
- Continuous Expert Assembly: Instance-Conditioned Low-Rank Residuals for All-in-One Image Restoration (NeurIPS 2026 · similarity 0.84)
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.
