AI Paper Atlas

Research map / Image generation / Super-resolution

Toward Accurate Post-Training Quantization for Image Super Resolution

Zhijun Tu, Jie Hu, Hanting Chen, Yunhe Wang

CVPR 2023 · Image generation › Super-resolution · approach: Perceptual and real-world SR · 27 citations (OpenAlex)

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 changedWhereMeasuredEffect
Use DBDC and PaC for 4-bit PTQ instead of MinMax PTQ (ablation)EDSR ×4 · Urban100PSNR (dB): 22.871 -> 25.556Improved
Add density-based dual clipping (ablation)EDSR ×4, 4-bit MinMax PTQ · Urban100PSNR (dB): 22.871 -> 25.312Improved
Add pixel-aware calibration (ablation)EDSR ×4, 4-bit MinMax PTQ · Urban100PSNR (dB): 22.871 -> 24.116Improved
Add PaC after DBDC (ablation)EDSR ×4, 4-bit PTQ · Urban100PSNR (dB): 25.312 -> 25.556Improved
Add DBDC after PaC (ablation)EDSR ×4, 4-bit PTQ · Urban100PSNR (dB): 24.116 -> 25.556Improved
Use the proposed PTQ method rather than the strongest listed baseline at each bit-widthEDSR ×4 · Urban100PSNR (dB): 26.560 -> 26.567; 25.890 -> 26.382; 22.871 -> 25.556Improved
Use the proposed PTQ method rather than the strongest listed baseline at each bit-widthSRResNet ×4 · Urban100PSNR (dB): 26.111 -> 26.191; 25.340 -> 26.011; 23.042 -> 25.133Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy