Research map / Video / Video segmentation and tracking
Towards Video Anomaly Detection from Event Streams: A Baseline and Benchmark Datasets
Peng Wu, Yuting Yan, Guansong Pang, Yujia Sun, Qingsen Yan, Peng Wang, Yanning Zhang
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 |
|---|---|---|---|
| add event-density aware dynamic sampling (EDS) (ablation) | VadCLIP-based event model · UCF-Crime (simulated) | AUC: 73.01 -> 73.94 | Improved |
| add event-modulated distance-decay attention (EDA) on top of EDS (ablation) | EWAD · UCF-Crime (simulated) | AUC: 73.94 -> 74.17 | Improved |
| add RGB-to-event knowledge distillation (B-KD + M-KD) alone (ablation) | event baseline · UCF-Crime (simulated) | AUC: 73.01 -> 74.45 | Improved |
| add multi-class distillation to binary distillation (ablation) | EWAD (EDS+EDA+B-KD) · UCF-Crime (simulated) | AUC: 74.53 -> 76.55 | Improved |
| add binary distillation to multi-class distillation (ablation) | EWAD (EDS+EDA+M-KD) · UCF-Crime (simulated) | AUC: 74.58 -> 76.55 | Improved |
| combine all components (ablation) | VadCLIP baseline · UCF-Crime (simulated) | AUC: 73.01 -> 76.55 | Improved |
| use EWAD instead of previous SNN SOTA (MSF) | EWAD · UCF-Crime (real event camera) | AUC: 65.01 -> 72.36 | Improved |
| use training-free event-based localization instead of RGB methods | EWAD · UCF-Crime localization | TIoU: 13.28 vs 22.05 (VadCLIP RGB) | Mixed |
| use EWAD on UBnormal versus VadCLIP | EWAD · UBnormal (simulated) | AUC: 53.19 -> 58.30 | Improved |
Closest papers by meaning
- Pistachio: Towards Synthetic, Balanced, and Long-Form Video Anomaly Benchmarks (ECCV 2026 · similarity 0.90)
- VADTree: Explainable Training-Free Video Anomaly Detection via Hierarchical Granularity-Aware Tree (NeurIPS 2025 · similarity 0.88)
- MoniTor: Exploiting Large Language Models with Instruction for Online Video Anomaly Detection (NeurIPS 2025 · similarity 0.88)
- Learning Event Completeness for Weakly Supervised Video Anomaly Detection (ICML 2025 · similarity 0.88)
- Advancing Video Anomaly Detection: A Concise Review and a New Dataset (NeurIPS 2024 · similarity 0.88)
- Just Dance with pi! A Poly-modal Inductor for Weakly-supervised Video Anomaly Detection (CVPR 2025 · similarity 0.87)
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.
