ReVi adapter enables off-the-shelf vision models to localize image manipulations by separating and enhancing manipulation cues from semantic features without full model retraining.
Imd2020: A large-scale annotated dataset tailored for de- tecting manipulated images
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Off-the-shelf Vision Models Benefit Image Manipulation Localization
ReVi adapter enables off-the-shelf vision models to localize image manipulations by separating and enhancing manipulation cues from semantic features without full model retraining.