AI Paper Atlas

Research map / Image generation / Image restoration

Deep Boosting for Image Denoising

Chang Chen, Zhiwei Xiong, Xinmei Tian, Feng Wu

ECCV 2018 · Image generation › Image restoration · approach: Self-supervised image denoising · 117 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
remove the subtraction of the previous estimate (no identical subtraction as in SOS) (ablation)3-stage DBF · grey denoisingPSNR: +0.12 dBImproved
increase boosting stages from 1 to 5 (ablation)DBF · Set12 sigma=50PSNR: 0.30 dB gainImproved
greedy training instead of joint training (ablation)3-stage DBF · grey denoisingPSNRNo clear effect
use deeper thinner plain network (PN2) instead of PN at equal parameters (ablation)PN · Set12 sigma=50PSNRImproved
put PN2 into a 2-stage boosting framework (ablation)PN2 · Set12 sigma=50PSNRNo clear effect
add dense connection (ablation)PN2 · Set12 sigma=50convergence and PSNRImproved
add dilated convolution (ablation)DN · Set12 sigma=50PSNRImproved
add path-widening fusion (ablation)DDN · Set12 sigma=50PSNRImproved
double the number of filters in each layer (W) (ablation)DDFN · Set12 sigma=50PSNRImproved
use DDFN-x5W boosting instead of BM3D with SOS boostingBM3D-SOS · Set5PSNR: +0.77 dB (sigma=25), +0.92 dB (sigma=50)Improved
apply SOS boosting to classic denoisersBM3D · Set5PSNR: 31.58 -> 31.61Improved
use DDFN-x5W instead of DnCNNDnCNN · Set12 sigma=50PSNR: +0.28 dBImproved

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