Research map / Image generation / Super-resolution
Learning Series-Parallel Lookup Tables for Efficient Image Super-Resolution
Cheng Ma, Jingyi Zhang, Jie Zhou, Jiwen Lu
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 parallel MSB/LSB branch instead of single branch with tail-layer interpolation (ablation) | SPLUT-M · Set5 x4 | PSNR: 27.24 -> 30.23 | Improved |
| add parallel MSB/LSB branch instead of single branch with tail-layer interpolation (ablation) | SPLUT-M · Set14 x4 | PSNR: 25.30 -> 27.32 | Improved |
| add interpolation (tail-layer or first-layer) to one-branch model (ablation) | SPLUT-M · Set5 x4 | PSNR: 27.24 -> 27.24 | No clear effect |
| add horizontal and vertical aggregation modules (ablation) | SPLUT-M · Set5 x4 | PSNR: 29.64 -> 30.23 | Improved |
| add horizontal and vertical aggregation modules (ablation) | SPLUT-M · Urban100 x4 | PSNR: 23.50 -> 24.21 | Improved |
| add horizontal and vertical aggregation modules (ablation) | SPLUT-M · Manga109 x4 | PSNR: 25.72 -> 27.20 | Improved |
| add a second query block (depth, SPLUT1-2 to SPLUT1-2-2) (ablation) | SPLUT-M · Set5 x4 | PSNR: 29.77 -> 30.23 | Improved |
| add a second query block (SPLUT1-4 to SPLUT1-4-4) (ablation) | SPLUT-L · Set5 x4 | PSNR: 30.01 -> 30.52 | Improved |
| widen layer (SPLUT1-2 to SPLUT1-4) (ablation) | SPLUT-M · Set5 x4 | PSNR: 29.77 -> 30.01 | Improved |
| add depth rather than width at equal LUT count (SPLUT1-2-2 vs SPLUT1-4) (ablation) | SPLUT-M · Set5 x4 | PSNR: 30.01 -> 30.23 | Improved |
| increase quantization precision vf from 8 to 16 (ablation) | SPLUT-M · Set5 x4 | PSNR: 29.95 -> 30.23 | Improved |
| increase quantization precision vf from 16 to 24 (ablation) | SPLUT-M · Set5 x4 | PSNR: 30.23 -> 30.25 | No clear effect |
Closest papers by meaning
- Learning Extensible Series-Parallel Lookup Tables for Efficient Image Super-Resolution (IJCV 2025 · similarity 0.91)
- Efficient Look-Up Table from Expanded Convolutional Network for Accelerating Image Super-resolution (AAAI 2024 · similarity 0.90)
- TinyLUT: Tiny Look-Up Table for Efficient Image Restoration at the Edge (NeurIPS 2024 · similarity 0.85)
- Pan-LUT: Efficient Pan-sharpening via Learnable Look-Up Tables (NeurIPS 2025 · similarity 0.85)
- Recursive Generalization Transformer for Image Super-Resolution (ICLR 2024 · similarity 0.85)
- GSDD: Generative Space Dataset Distillation for Image Super-resolution (AAAI 2024 · 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.
