atlas

Research map / Image generation / Image restoration

Region-Adaptive Dense Network for Efficient Motion Deblurring

AAAI 2020 · Image generation › Image restoration · approach: Image deblurring · 115 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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 deformable (spatially adaptive offset) units in dense decoder modules instead of ordinary convs (ablation)RADN · GoPro testPSNR: 31.58 -> 31.76Helped
add self-attention (SA) module (ablation)RADN4 · GoPro testPSNR: 31.42 -> 31.58Helped
use dense modules instead of bottleneck blocks (ablation)RADN · GoPro testPSNR: 31.17 -> 31.42Helped
pretrain encoder (DenseNet) (ablation)RADN2 · GoPro testPSNR: 30.77 -> 31.17Helped
replace 6 convs with deformable layers (ablation)residual baseline · GoPro testPSNR: 29.8 -> 31.05Helped
add SA module to plain CNN (ablation)residual baseline · GoPro testPSNR: 29.8 -> 30.54Helped
add SA module on top of 6 deformable layers (ablation)residual baseline · GoPro testPSNR: 31.05 -> 31.13Helped
add geometric self-ensemble at test timeRADN(Final) · GoPro testPSNR: 31.76 -> 32.15Helped
RADN vs Stack(4)-DMPHNmotion deblurring · GoPro testruntime: 700 -> 38Helped

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