Research map / Video / Video segmentation and tracking
A Hybrid Video Anomaly Detection Framework via Memory-Augmented Flow Reconstruction and Flow-Guided Frame Prediction
Zhian Liu, Yongwei Nie, Chengjiang Long, Qing Zhang, Guiqing Li
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 frame prediction to flow reconstruction (ablation) | HF2-VAD · UCSD Ped2 | AUROC (%): 98.8 -> 99.3 | Improved |
| add frame prediction to flow reconstruction (ablation) | HF2-VAD · CUHK Avenue | AUROC (%): 86.8 -> 91.1 | Improved |
| add frame prediction to flow reconstruction (ablation) | HF2-VAD · ShanghaiTech | AUROC (%): 73.1 -> 76.2 | Improved |
| add flow reconstruction to frame prediction (ablation) | HF2-VAD · UCSD Ped2 | AUROC (%): 94.5 -> 99.3 | Improved |
| add flow reconstruction to frame prediction (ablation) | HF2-VAD · CUHK Avenue | AUROC (%): 90.2 -> 91.1 | Improved |
| add flow reconstruction to frame prediction (ablation) | HF2-VAD · ShanghaiTech | AUROC (%): 76.0 -> 76.2 | Improved |
| use ML-MemAE-SC with 3 memories instead of MemAE (ablation) | Flow reconstruction · UCSD Ped2 | AUROC (%): 96.27 -> 98.81 | Improved |
| use CVAE with flow condition instead of VAE for prediction (ablation) | Frame prediction · UCSD Ped2 | AUROC (%): 89.96 -> 94.48 | Improved |
| combine ML-MemAE-SC (3 memories) with CVAE instead of reconstruction only (ablation) | Flow reconstruction · UCSD Ped2 | AUROC (%): 98.81 -> 99.31 | Improved |
| combine ML-MemAE-SC (3 memories) with CVAE instead of prediction only (ablation) | CVAE prediction · UCSD Ped2 | AUROC (%): 94.48 -> 99.31 | Improved |
| feed all four reconstructed flows to CVAE instead of one (ablation) | HF2-VAD · UCSD Ped2 | AUROC (%): 98.70 -> 99.31 | Improved |
| use deterministic posterior-mean sampling instead of stochastic sampling at test time (ablation) | CVAE · UCSD Ped2 | AUROC (%): 99.3078 vs 99.3065 to 99.3089 | No clear effect |
Closest papers by meaning
- Noise-Resistant Video Anomaly Detection via RGB Error-Guided Multiscale Predictive Coding and Dynamic Memory (CVPR 2025 · similarity 0.87)
- Deep Anomaly Discovery From Unlabeled Videos via Normality Advantage and Self-Paced Refinement (CVPR 2022 · similarity 0.86)
- Video Event Restoration Based on Keyframes for Video Anomaly Detection (CVPR 2023 · similarity 0.86)
- Self-Distilled Masked Auto-Encoders are Efficient Video Anomaly Detectors (CVPR 2024 · similarity 0.85)
- VADTree: Explainable Training-Free Video Anomaly Detection via Hierarchical Granularity-Aware Tree (NeurIPS 2025 · 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.
