Training-Free From survey PDF — verify links
SUVAD
SUVAD — Semantic Understanding
ICASSP · 2025Human-reviewed summary status
Survey summary
SUVAD is a training-free MLLM-based method that generates detailed textual descriptions of video content for semantic understanding, then detects anomalies directly via an LLM from those descriptions, with techniques to mitigate MLLM hallucination and support flexible, scene-adaptable anomaly definitions.
Taxonomy placement
Training-Free
Training-FreeSemantic Understanding
Method characteristics
Training-FreeMLLMEvent listsFlexible definitions