AD-FM combines multi-stage reasoning with localization-aware rewards to fine-tune MLLMs for anomaly detection, improving average accuracy by about 22 percentage points over the base model.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
AD-FM: Multimodal LLMs for Anomaly Detection via Multi-Stage Reasoning and Fine-Grained Reward Optimization
AD-FM combines multi-stage reasoning with localization-aware rewards to fine-tune MLLMs for anomaly detection, improving average accuracy by about 22 percentage points over the base model.