Research map / Video / Video segmentation and tracking
Dance With Self-Attention: A New Look of Conditional Random Fields on Anomaly Detection in Videos
Didik Purwanto, Yie-Tarng Chen, Wen-Hsien Fang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use relation-aware feature extractor instead of TRN (ablation) | weakly supervised anomaly detector · UCF-Crime | AUC: 81.52 -> 82.43 | Improved |
| use relation-aware feature extractor instead of TRN (ablation) | weakly supervised anomaly detector · ShanghaiTech | AUC: 95.01 -> 95.43 | Improved |
| add self-attention CRF (ablation) | relation-aware network · UCF-Crime | AUC: 82.43 -> 83.89 | Improved |
| add self-attention CRF (ablation) | relation-aware network · ShanghaiTech | AUC: 95.43 -> 96.67 | Improved |
| add contrastive loss (ablation) | relation-aware + self-attention CRF · UCF-Crime | AUC: 83.89 -> 85.00 | Improved |
| add contrastive loss (ablation) | relation-aware + self-attention CRF · ShanghaiTech | AUC: 96.67 -> 96.85 | Improved |
| use cliques {1,2,4,6,8} instead of {1} in self-attention (ablation) | self-attention CRF · UCF-Crime | AUC: 83.03 -> 85.00 | Improved |
| use cliques {1,2,4,6,8} instead of {1} in self-attention (ablation) | self-attention CRF · ShanghaiTech | AUC: 95.32 -> 96.85 | Improved |
| add cliques 6 and 8 to {1,2,4} (ablation) | self-attention CRF · UCF-Crime | AUC: 84.06 -> 85.00 | Improved |
Closest papers by meaning
- Anomaly Detection in Video via Self-Supervised and Multi-Task Learning (CVPR 2021 · similarity 0.90)
- Look Around for Anomalies: Weakly-Supervised Anomaly Detection via Context-Motion Relational Learning (CVPR 2023 · similarity 0.89)
- Dual Memory Units with Uncertainty Regulation for Weakly Supervised Video Anomaly Detection (AAAI 2023 · similarity 0.88)
- MIST: Multiple Instance Self-Training Framework for Video Anomaly Detection (CVPR 2021 · similarity 0.88)
- UCF-Crime-DVS: A Novel Event-Based Dataset for Video Anomaly Detection with Spiking Neural Networks (AAAI 2025 · similarity 0.87)
- Prompt-Enhanced Multiple Instance Learning for Weakly Supervised Video Anomaly Detection (CVPR 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
