Research map / Video / Video segmentation and tracking
EVAL: Explainable Video Anomaly Localization
Ashish Singh, Michael J. Jones, Erik G. Learned-Miller
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 EVAL instead of the strongest listed baseline | video anomaly localization · CUHK Avenue | RBDC AUC: 65.99 -> 68.2 | Improved |
| Use EVAL instead of the strongest listed baseline | video anomaly localization · CUHK Avenue | TBDC AUC: 89.28 -> 87.56 | Hurt |
| Use EVAL instead of the strongest listed baseline | video anomaly localization · CUHK Avenue | Frame AUC: 92.9 -> 86.02 | Hurt |
| Use EVAL instead of the strongest listed baseline | video anomaly localization · ShanghaiTech | RBDC AUC: 45.45 -> 59.21 | Improved |
| Use EVAL instead of the strongest listed baseline | video anomaly localization · ShanghaiTech | TBDC AUC: 84.50 -> 89.44 | Improved |
| Use EVAL instead of the strongest listed baseline | video anomaly localization · ShanghaiTech | Frame AUC: 90.02 -> 76.63 | Hurt |
| Use EVAL instead of the strongest listed baseline | video anomaly localization · Street Scene | RBDC AUC: 21.0 -> 24.26 | Improved |
| Use EVAL instead of the strongest listed baseline | video anomaly localization · Street Scene | TBDC AUC: 53.0 -> 64.5 | Improved |
| Combine appearance and motion features instead of using either alone (ablation) | EVAL · UCSD Ped1, UCSD Ped2, CUHK Avenue, Street Scene | RBDC and TBDC AUC | Mixed |
| Replace EVAL appearance embeddings with ImageNet pre-classification features (ablation) | EVAL · UCSD Ped1 | RBDC AUC: 29.6 -> 25.148 | Hurt |
| Replace EVAL appearance embeddings with ImageNet pre-classification features (ablation) | EVAL · UCSD Ped1 | TBDC AUC: 64.8 -> 44.63 | Hurt |
| Replace EVAL appearance embeddings with ImageNet pre-classification features (ablation) | EVAL · UCSD Ped2 | RBDC AUC: 77.7 -> 64.67 | Hurt |
Closest papers by meaning
- Track Any Anomalous Object:A Granular Video Anomaly Detection Pipeline (CVPR 2025 · similarity 0.88)
- HAWK: Learning to Understand Open-World Video Anomalies (NeurIPS 2024 · similarity 0.87)
- A Unified Reasoning Framework for Holistic Zero-Shot Video Anomaly Analysis (NeurIPS 2025 · similarity 0.87)
- VERA: Explainable Video Anomaly Detection via Verbalized Learning of Vision-Language Models (CVPR 2025 · similarity 0.87)
- Video Anomaly Detection with Motion and Appearance Guided Patch Diffusion Model (AAAI 2025 · similarity 0.87)
- Holmes-VAU: Towards Long-term Video Anomaly Understanding at Any Granularity (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.
