atlas

Research map / Image generation / Image restoration

Robust Unsupervised StyleGAN Image Restoration

CVPR 2023 · Image generation › Image restoration · approach: Priors for restoration networks · 22 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 the proposed three-phase latent extension, normalized gradient descent, and multiresolution loss with one fixed hyperparameter setStyleGAN2-ADA restoration · Upsampling, denoising, deartifacting, and inpainting at XS–XL degradation levelsaccuracy (LPIPS)Helped
Use the proposed method rather than DDRMStyleGAN2-ADA restoration · Inexact upsampling, nonlinear denoising, and uncroppingaccuracy (LPIPS): 0.559 vs 0.539; 0.530 vs 0.450; 0.468 vs 0.418Helped
Use the proposed method rather than DDRMStyleGAN2-ADA restoration · Exact upsampling and linear denoisingaccuracy (LPIPS): 0.412 vs 0.443; 0.336 vs 0.452Hurt
Use the proposed method rather than DDRMStyleGAN2-ADA restoration · All DDRM comparison scenariosrealism (pFID): 44.75 vs 18.92; 67.06 vs 32.81; 43.28 vs 19.25; 156.1 vs 19.06; 26.12 vs 25.71; 34.33 vs 17.645Helped
Compose multiple degradation operators without retuning hyperparametersProposed StyleGAN restoration method · Medium-level composed degradationsaccuracy (LPIPS)Helped

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