Research map / Image generation / Super-resolution
SPR$^2$Q: Static Priority-based Rectifier Routing Quantization for Image Super-Resolution
Jingwei Xin, Wenhao Li, Nannan Wang, Jie Li, Xinbo Gao
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 pre-quantization fused low-rank rectifier (PQFR) (ablation) | MambaIRv2-light 4-bit · Set5 x2 / Urban100 x2 | PSNR: Set5 37.20 -> 37.44; Urban100 30.69 -> 31.25 | Improved |
| add rectifier group training (RGT) (ablation) | SPR2Q with PQFR · Set5 x2 | PSNR: 37.44 -> 37.60 | Improved |
| add rectifier group training (RGT) (ablation) | SPR2Q with PQFR · Urban100 x2 | PSNR: 31.25 -> 31.24 | No clear effect |
| add offline static routing calibration (OSRC) (ablation) | SPR2Q with PQFR and RGT · Set5 x2 / Urban100 x2 | PSNR: Set5 37.60 -> 37.72; Urban100 31.24 -> 31.53 | Improved |
| increase rectifier rank from 8 to 16 (ablation) | SPR2Q · Set5 x2 | PSNR: 37.72 -> 37.74 | No clear effect |
| increase rectifier rank from 2 to 8 (ablation) | SPR2Q · Set5 x2 | PSNR: 37.63 -> 37.72 | Improved |
| increase rectifier group size from 2 to 4 (ablation) | SPR2Q · Set5 x2 | PSNR: 37.50 -> 37.72 | Improved |
| increase rectifier group size from 4 to 8 (ablation) | SPR2Q · Set5 x2 | PSNR: 37.72 -> 37.82 | Improved |
| use SPR2Q instead of Mamba PTQ baselines at 4-bit | PTQ4VM/Quamba/MambaQuant · Urban100 x2 | PSNR: 30.54 -> 31.53 | Improved |
| use SPR2Q instead of Mamba PTQ baselines at 2-bit | Quamba · Set5 x2 | PSNR: 34.66 -> 35.97 | Improved |
| apply SPR2Q instead of ViT PTQ methods to SwinIR-light at 2-bit | APHQ-ViT · Set5 x2 | PSNR: 36.14 -> 37.28 | Improved |
| quantize from 2-bit to 1-bit (ablation) | SPR2Q · Set5 x2 | PSNR: 35.97 -> 34.82 | Hurt |
Later papers that compared with this method
1 later accepted paper reports SPR2Q as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Hardware-Friendly Token-Group Activation Quantization for Low-Bit Mamba Super-Resolution (NeurIPS 2026 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- QuantSR: Accurate Low-bit Quantization for Efficient Image Super-Resolution (NeurIPS 2023 · similarity 0.93)
- 2DQuant: Low-bit Post-Training Quantization for Image Super-Resolution (NeurIPS 2024 · similarity 0.90)
- AdaBM: On-the-Fly Adaptive Bit Mapping for Image Super-Resolution (CVPR 2024 · similarity 0.88)
- AdaFormer: Efficient Transformer with Adaptive Token Sparsification for Image Super-resolution (AAAI 2024 · similarity 0.87)
- Thinking in Granularity: Dynamic Quantization for Image Super-Resolution by Intriguing Multi-Granularity Clues (AAAI 2025 · similarity 0.87)
- PassionSR: Post-Training Quantization with Adaptive Scale in One-Step Diffusion based Image Super-Resolution (CVPR 2025 · similarity 0.87)
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.
