COVE predicts an image's true context before judging caption veracity, improving real-world out-of-context detection and providing a reusable context artifact for human verifiers.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
COVE: COntext and VEracity prediction for out-of-context images
COVE predicts an image's true context before judging caption veracity, improving real-world out-of-context detection and providing a reusable context artifact for human verifiers.