Classifying from spatially aggregated DINO patch tokens instead of a single CLS token improves cross-dataset AI-generated image detection by about three percentage points average balanced accuracy.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
PatchHead: Learning Spatial Patch Evidence for Generalizable AI-Generated Image Detection
Classifying from spatially aggregated DINO patch tokens instead of a single CLS token improves cross-dataset AI-generated image detection by about three percentage points average balanced accuracy.