LOGER ensembles heterogeneous global vision models with selective local patch aggregation via multiple instance learning to achieve robust deepfake detection across varied manipulations and degradations.
Veritas: Generalizable deepfake detection via pattern-aware reasoning
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
-
LOGER: Local--Global Ensemble for Robust Deepfake Detection in the Wild
LOGER ensembles heterogeneous global vision models with selective local patch aggregation via multiple instance learning to achieve robust deepfake detection across varied manipulations and degradations.