AI Paper Atlas

Research map / Image generation / Editing and style transfer

Deep Saliency Prior for Reducing Visual Distraction

Kfir Aberman, Junfeng He, Yossi Gandelsman, Inbar Mosseri, David E. Jacobs, Kai Kohlhoff, Yael Pritch, Michael Rubinstein

CVPR 2022 · Image generation › Editing and style transfer · approach: Localized and attribute-specific editing · 23 citations (OpenAlex)

Check 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
Use saliency-guided editing instead of Look-here!Distracting regions in images · Perceptual study, predicted saliencysaliency reduction: 25.8% vs 43.1% (Recolor), 92.9% (Warp), and 53.3% (ConvNet)Improved
Use saliency-guided warping rather than Look-here!Distracting regions in images · User preference, Warp comparisonuser preference: 9.4% vs 84.4Improved
Use saliency-guided recoloring rather than Look-here!Distracting regions in images · User preference, Recolor comparisonuser preference: 31.3% vs 62.5%Improved
Use the saliency-guided ConvNet effect rather than Look-here!Distracting regions in images · User preference, ConvNet comparisonuser preference: 18.8% vs 75%Improved
Apply saliency-guided editing with the tested operatorsMarked regions in images · User gaze study, 31 imagesgaze saliency within the mask: t=-4.5; p=6×10^{-5}Improved
Apply saliency-guided editing with the tested operatorsMarked regions in images · User gaze study, 31 imagesconsecutive gaze duration within the mask: t=-3.3; p=2.2×10^{-5}Improved
Apply saliency-guided editing with the tested operatorsMarked regions in images · User gaze study, 31 imagesfirst gaze time: t=3.5; p=1.2×10^{-3}Improved
Use saliency-guided warping instead of recoloring, ConvNet, or GANMarked regions in the real-gaze study · Perceptual study, real gazegaze saliency reduction: 43.1% (Recolor), 92.9% (Warp), 53.3% (ConvNet), 34.8% (GAN)Mixed
Use the proposed edits instead of leaving the image uneditedImages in the realism study · User realism studyimages marked realistic: 85% -> 78%Hurt

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