Research map / Video / Video segmentation and tracking
Future Frame Prediction for Anomaly Detection – A New Baseline
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 future-frame prediction with appearance and motion constraints for anomaly detection | video anomaly detection · CUHK Avenue | frame-level AUC: 84.6% -> 85.1% | Helped |
| Use future-frame prediction with appearance and motion constraints for anomaly detection | video anomaly detection · UCSD Ped1 | frame-level AUC: 97.4% -> 83.1% | Hurt |
| Use future-frame prediction with appearance and motion constraints for anomaly detection | video anomaly detection · UCSD Ped2 | frame-level AUC: 94.1% -> 95.4% | Helped |
| Use future-frame prediction with appearance and motion constraints for anomaly detection | video anomaly detection · ShanghaiTech | frame-level AUC: 68.0% -> 72.8% | Helped |
| Use U-Net rather than Beyond-MSE for the prediction network (ablation) | future-frame prediction anomaly detection · UCSD Ped1 | AUC: 75.8% -> 81.8% | Helped |
| Use U-Net rather than Beyond-MSE for the prediction network (ablation) | future-frame prediction anomaly detection · UCSD Ped2 | AUC: 88.5% -> 93.5% | Helped |
| Add the optical-flow motion constraint (ablation) | U-Net future-frame predictor · UCSD Ped1 | AUC: 81.8% -> 83.1% | Helped |
| Add the optical-flow motion constraint (ablation) | U-Net future-frame predictor · UCSD Ped2 | AUC: 93.5% -> 95.4% | Helped |
| Add gradient loss to intensity loss (ablation) | future-frame predictor · CUHK Avenue | AUC: 82.0% -> 82.6% | Helped |
| Add adversarial loss to intensity and gradient losses (ablation) | future-frame predictor · CUHK Avenue | AUC: 82.6% -> 83.7% | Helped |
| Add optical-flow loss to intensity, gradient, and adversarial losses (ablation) | future-frame predictor · CUHK Avenue | AUC: 83.7% -> 85.1% | Helped |
| Add the optical-flow motion constraint (ablation) | future-frame predictor · UCSD Ped1 optical-flow maps | MSE: 8.26 -> 7.51 | Helped |
Closest papers by meaning
- Video Anomaly Detection with Motion and Appearance Guided Patch Diffusion Model (AAAI 2025 · similarity 0.89)
- Holmes-VAU: Towards Long-term Video Anomaly Understanding at Any Granularity (CVPR 2025 · similarity 0.88)
- Noise-Resistant Video Anomaly Detection via RGB Error-Guided Multiscale Predictive Coding and Dynamic Memory (CVPR 2025 · similarity 0.87)
- Dual Conditioned Motion Diffusion for Pose-Based Video Anomaly Detection (AAAI 2025 · similarity 0.87)
- FrameShield: Adversarially Robust Video Anomaly Detection (NeurIPS 2025 · similarity 0.87)
- Track Any Anomalous Object:A Granular Video Anomaly Detection Pipeline (CVPR 2025 · 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.