AI Paper Atlas

Research map / Image generation / Image restoration

Pyramid Attention Network for Image Restoration

Yiqun Mei, Yuchen Fan, Yulun Zhang, Jiahui Yu, Yuqian Zhou, Ding Liu, Yun Fu, Thomas S. Huang, Humphrey Shi

IJCV 2023 · Image generation › Image restoration · approach: Priors for restoration networks · 253 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
add single-scale non-local attention to the ResNet baseline (ablation)ResNet 16 blocks · Urban100 denoising sigma=30PSNR: 30.86 -> 31.14Improved
replace non-local attention with pyramid attention (ablation)ResNet 16 blocks · Urban100 denoising sigma=30PSNR: 31.14 -> 31.29; 0.43 dB over the baseline and 0.15 dB over classical non-localImproved
use 3x3 block-wise matching instead of pixel-wise matching (ablation)non-local attention · Urban100 denoising sigma=30PSNR: 31.14 dB to 31.21 dBImproved
increase pyramid from 1 level to 5 levels (ablation)pyramid attention · Urban100 denoising sigma=30PSNR: 31.21 -> 31.29Improved
insert pyramid attention at the middle position only (ablation)ResNet · Urban100 denoising sigma=30PSNR: 30.86 -> 31.29Improved
insert pyramid attention at the pre position only (ablation)ResNet · Urban100 denoising sigma=30PSNR: 30.86 -> 31.07Improved
insert pyramid attention at the post position only (ablation)ResNet · Urban100 denoising sigma=30PSNR: 30.86 -> 31.18Improved
insert pyramid attention at pre, mid and post positions together (ablation)ResNet · Urban100 denoising sigma=30PSNR: 31.29 -> 31.48Improved
add one pyramid attention block to a plain ResNet backbone instead of RNAN's multiple non-local blocksPANet · Urban100 denoisingPSNR: more than 0.3 dB improvements over RNAN on all noisy levelsImproved
add pyramid attention to EDSR for super resolutionEDSR · Urban100 x4 SRPSNR: 26.64 -> 27.01Improved
use a small 8-block PANet-S instead of RNAN (1LB+1NLB)PANet-S · Urban100 denoisingPSNR: 28.36 -> 28.80 with 665K vs 1494K parametersImproved
add pyramid attention to PANet for demosaicing versus RNANPANet · Urban100 demosaicingPSNR: 39.75 -> 40.50Improved

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