A new multimodal dataset and benchmark shows that intent-aware classification of AI-generated images is hard, with the best model reaching only 71.6% accuracy in the wild.
Texturecrop: Enhancing synthetic image detection through texture-based cropping
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
"Humor, Art, or Misinformation?": A Multimodal Dataset for Intent-Aware Synthetic Image Detection
A new multimodal dataset and benchmark shows that intent-aware classification of AI-generated images is hard, with the best model reaching only 71.6% accuracy in the wild.