AI Paper Atlas

Research map / Image generation / Image restoration

Recurrent Squeeze-and-Excitation Context Aggregation Net for Single Image Deraining

Xia Li, Jianlong Wu, Zhouchen Lin, Hong Liu, Hongbin Zha

ECCV 2018 · Image generation › Image restoration · approach: Dehazing and deraining · 991 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 squeeze-and-excitation blocks to CAN (SCAN vs CAN) (ablation)SCAN · Rain800PSNR: 22.45 -> 23.45Improved
add squeeze-and-excitation blocks to CAN (SCAN vs CAN) (ablation)SCAN · Rain100HPSNR: 22.93 -> 23.56Improved
remove batch normalization from SCAN (ablation)SCAN · Rain800PSNR: 23.11 -> 23.45Improved
remove batch normalization from SCAN (ablation)SCAN · Rain100HPSNR: 23.09 -> 23.56Improved
use exponentially increasing dilation instead of plain dilation 1 (ablation)SCAN/CAN · Rain800PSNR: 22.10 -> 22.45Improved
use exponentially increasing dilation instead of plain dilation 1 (ablation)SCAN/CAN · Rain100HPSNR: 21.46 -> 22.93Improved
add recurrent stages with ConvGRU full prediction on top of SCAN (ablation)RESCAN · Rain800PSNR: 23.45 -> 24.09Improved
add recurrent stages with ConvGRU full prediction on top of SCAN (ablation)RESCAN · Rain100HPSNR: 23.56 -> 26.45Improved
keep previous-stage states (recurrent) instead of iterative framework without states (ablation)SCAN · Rain100HPSNR: 22.81 -> 26.45Improved
use full prediction instead of additive prediction (ConvGRU) (ablation)RESCAN · Rain100HPSNR: 24.00 -> 26.45Improved
use full prediction instead of additive prediction (ConvGRU) (ablation)RESCAN · Rain800PSNR: 23.31 -> 24.18Improved
use ConvGRU instead of ConvRNN with full prediction (ablation)RESCAN · Rain100HPSNR: 23.44 -> 26.45Improved

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