Research map / Video / Video segmentation and tracking
CLAWS: Clustering Assisted Weakly Supervised Learning with Normalcy Suppression for Anomalous Event Detection
Muhammad Zaigham Zaheer, Arif Mahmood, Marcella Astrid, Seung-Ik Lee
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 random batch selection (RBS) to the backbone (ablation) | BBN · UCF-Crime | frame-level AUC (%): 69.50 -> 75.95 | Improved |
| add random batch selection (RBS) to the backbone (ablation) | BBN · ShanghaiTech | frame-level AUC (%): 67.50 -> 79.64 | Improved |
| add normalcy suppression modules (NSM) to BBN+RBS (ablation) | BBN + RBS · ShanghaiTech | frame-level AUC (%): 79.64 -> 87.76 | Improved |
| add both NSMs to BBN+RBS (ablation) | BBN + RBS · UCF-Crime | frame-level AUC (%): 75.95 -> 80.94 | Improved |
| add NSM-1 to BBN+RBS (ablation) | BBN + RBS · UCF-Crime | frame-level AUC (%): 75.95 -> 78.60 | Improved |
| add NSM-2 on top of NSM-1 (ablation) | BBN + RBS + NSM-1 · UCF-Crime | frame-level AUC (%): 78.60 -> 80.94 | Improved |
| add sparsity and temporal consistency losses (ablation) | BBN + RBS + NSM · UCF-Crime | frame-level AUC (%): 80.94 -> 81.53 | Improved |
| add clustering distance loss (ablation) | CLAWS Net w/o clustering loss · UCF-Crime | frame-level AUC (%): 81.53 -> 83.03 | Improved |
| keep RBS in the full model (removing it lowers AUC) (ablation) | CLAWS Net · UCF-Crime | frame-level AUC (%): 80.23 -> 83.03 | Improved |
| keep NSM-1 in the full model (ablation) | CLAWS Net · UCF-Crime | frame-level AUC (%): 77.39 -> 83.03 | Improved |
| keep NSM-2 in the full model (ablation) | CLAWS Net · UCF-Crime | frame-level AUC (%): 79.78 -> 83.03 | Improved |
| keep both NSMs in the full model (ablation) | CLAWS Net · UCF-Crime | frame-level AUC (%): 76.81 -> 83.03 | Improved |
Closest papers by meaning
- Exploiting Completeness and Uncertainty of Pseudo Labels for Weakly Supervised Video Anomaly Detection (CVPR 2023 · similarity 0.90)
- The Road Less Seen: Segment Exploration for Weakly Supervised Video Anomaly Detection (CVPR 2026 · similarity 0.88)
- Learning Event Completeness for Weakly Supervised Video Anomaly Detection (ICML 2025 · similarity 0.88)
- Prompt-Enhanced Multiple Instance Learning for Weakly Supervised Video Anomaly Detection (CVPR 2024 · similarity 0.86)
- Self-Distilled Masked Auto-Encoders are Efficient Video Anomaly Detectors (CVPR 2024 · similarity 0.85)
- Text Prompt with Normality Guidance for Weakly Supervised Video Anomaly Detection (CVPR 2024 · similarity 0.85)
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.
