Research map / Image generation / Super-resolution
Towards Robust Full Low-bit Quantization of Super Resolution Networks
Denis S. Makhov, Irina Zhelavskaya, Ruslan Ostapets, Dehua Song, Kirill Solodskikh
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 |
|---|---|---|---|
| wrap quantized SR network with finite-difference operator and regularized PDE solver | EDSR x2 W4A4 · DIV2K | PSNR (dB): 30.23 -> 33.98 | Improved |
| wrap quantized SR network with operator and PDE solver | RFDN x2 W4A4 · DIV2K | PSNR (dB): 30.09 -> 33.76 | Improved |
| use regularized PDE solver instead of inverse filtering (ablation) | SRCNN x3 · Set5 | PSNR (dB): 32.27 -> 32.65 | Improved |
| use regularized PDE solver instead of inverse filtering (ablation) | SRCNN x3 · Urban100 | LPIPS: 0.240 -> 0.213 | Improved |
| apply method at lower bitwidths for EDSR x2 (ablation) | Ours vs LSQ+ · EDSR x2, 2-bit | PSNR | Improved |
| compare with DAQ and CADYQ for full 4-bit | EDSR x2 Urban100 · Urban100 | PSNR (dB): 30.84 -> 31.14 over CADYQ | Improved |
| add differential operator and PDE solver to quantized network | EDSR x2 W4A4 · BitOps | BitOps (G): 5.63 -> 6.41 | Hurt |
| add method to SRCNN x2 W4A4 versus FP | SRCNN x2 · Set5 | PSNR (dB): 34.23 -> 37.18 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CADYQ — CADyQ: Content-Aware Dynamic Quantization for Image Super-Resolution (ECCV 2022) · 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 0.91)
- Condition Number Based Low-Bit Quantization for Image Super-Resolution (ICML 2026 · similarity 0.89)
- Robust Quantization: One Model to Rule Them All (NeurIPS 2020 · similarity 0.89)
- AdaBM: On-the-Fly Adaptive Bit Mapping for Image Super-Resolution (CVPR 2024 · similarity 0.88)
- 2DQuant: Low-bit Post-Training Quantization for Image Super-Resolution (NeurIPS 2024 · similarity 0.88)
- PassionSR: Post-Training Quantization with Adaptive Scale in One-Step Diffusion based Image Super-Resolution (CVPR 2025 · similarity 0.86)
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.
