AI Paper Atlas

Research map / Image generation / Image restoration

Multi-view Self-supervised Disentanglement for General Image Denoising

Hao Chen, Chenyuan Qu, Yu Zhang, Chen Chen, Jianbo Jiao

ICCV 2023 · Image generation › Image restoration · approach: Self-supervised image denoising · 11 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 Bernoulli manifold mixture loss (weight 0.025 vs 0.000) (ablation)MeD · CBSD68 Gaussian sigma=75PSNR: 23.78 -> 26.01Improved
raise manifold loss weight from 0.025 to 0.050 (ablation)MeD · CBSD68 Gaussian sigma=75PSNR: 26.01 -> 24.29Hurt
increase views from 2 to 4 (ablation)MeD · Noise Pool SpecklePSNR: 37.21 -> 37.47Improved
increase views from 2 to 4 (ablation)MeD · training costtraining time: 10x training timeHurt
pretrain with MeD instead of N2C (ablation)N2C fine-tuning · CBSD68 unseen noise averagePSNR: 33.50 -> 34.46Improved
pretrain with MeD instead of N2N (ablation)N2N fine-tuning · CBSD68 unseen noise averagePSNR: 33.59 -> 34.00Improved
pretrain with MeD instead of LIR (ablation)LIR fine-tuning · CBSD68 unseen noise averagePSNR: 28.38 -> 29.73Improved
train with random sigma instead of fixed sigma=25 (ablation)MeD · CBSD68 Gaussian sigma=75PSNR: 23.09 -> 25.40Improved
train MeD on Noise Pool rather than using supervised N2CMeD vs N2C · CBSD68 Noise Pool averagePSNR: 34.08 -> 36.02Improved
use MeD on low noise level with fixed sigma 25 trainingMeD vs DBD4 · CBSD68 sigma=15PSNR: 33.57 -> 33.11Hurt

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