Holmes-VAD: Towards Unbiased and Explainable Video Anomaly Detection via Multi-modal LLM.
Huaxin ZhangXiaohao XuXiang WangJialong ZuoChuchu HanXiaonan HuangChangxin GaoYuehuan WangNong SangPublished in: CoRR (2024)
Keyphrases
- multi modal
- anomaly detection
- video search
- semantic concepts
- intrusion detection
- network anomaly detection
- network traffic
- anomalous behavior
- detecting anomalies
- video data
- network intrusion detection
- multi modality
- detect anomalies
- video sequences
- multimedia
- multiple modalities
- negative selection algorithm
- video frames
- audio visual
- intrusion detection system
- one class support vector machines
- uni modal
- video content
- video shots
- unsupervised learning
- image processing
- cumulative sum
- activity recognition