Research map / Video / Video segmentation and tracking
MGFN: Magnitude-Contrastive Glance-and-Focus Network for Weakly-Supervised Video Anomaly Detection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 MGFN with I3D-RGB features instead of MSL (2022) with I3D-RGB | Weakly supervised video anomaly detection · UCF-Crime | AUC (%): 86.98 vs 85.30 | Helped |
| Use MGFN with VideoSwin-RGB features instead of MSL (2022) with VideoSwin-RGB | Weakly supervised video anomaly detection · UCF-Crime | AUC (%): 86.67 vs 85.62 | Helped |
| Use MGFN with I3D-RGB features instead of MSL (2022) with I3D-RGB | Weakly supervised video anomaly detection · XD-Violence | AP (%): 79.19 vs 78.28 | Helped |
| Use MGFN with VideoSwin-RGB features instead of MSL (2022) with VideoSwin-RGB | Weakly supervised video anomaly detection · XD-Violence | AP (%): 80.11 vs 78.58 | Helped |
| Use Glance followed by Focus rather than two cascaded Focus Blocks (ablation) | MGFN · UCF-Crime | AUC (%): 82.66 -> 85.80 | Helped |
| Use Glance followed by Focus rather than two cascaded Focus Blocks (ablation) | MGFN · XD-Violence | AP (%): 72.14 -> 78.50 | Helped |
| Place Glance after Focus rather than before Focus (ablation) | MGFN · UCF-Crime | AUC (%): 83.10 -> 85.80 | Helped |
| Place Glance after Focus rather than before Focus (ablation) | MGFN · XD-Violence | AP (%): 74.30 -> 78.50 | Helped |
| Use Glance followed by Focus rather than parallel Glance-Focus fusion (ablation) | MGFN · UCF-Crime | AUC (%): 85.05 -> 85.80 | Helped |
| Use Glance followed by Focus rather than parallel Glance-Focus fusion (ablation) | MGFN · XD-Violence | AP (%): 78.03 -> 78.50 | Helped |
| Add Feature Amplification Mechanism (ablation) | MGFN baseline without FAM or MC loss · UCF-Crime | AUC (%): 83.20 -> 84.85 | Helped |
| Add Feature Amplification Mechanism (ablation) | MGFN baseline without FAM or MC loss · XD-Violence | AP (%): 75.11 -> 76.13 | Helped |
Closest papers by meaning
- Learning Event Completeness for Weakly Supervised Video Anomaly Detection (ICML 2025 · similarity 0.86)
- VadCLIP: Adapting Vision-Language Models for Weakly Supervised Video Anomaly Detection (AAAI 2024 · similarity 0.86)
- Prompt-Enhanced Multiple Instance Learning for Weakly Supervised Video Anomaly Detection (CVPR 2024 · similarity 0.84)
- UCF-Crime-DVS: A Novel Event-Based Dataset for Video Anomaly Detection with Spiking Neural Networks (AAAI 2025 · similarity 0.84)
- Multi-Scale Video Anomaly Detection by Multi-Grained Spatio-Temporal Representation Learning (CVPR 2024 · similarity 0.84)
- Just Dance with pi! A Poly-modal Inductor for Weakly-supervised Video Anomaly Detection (CVPR 2025 · similarity 0.84)
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.