Research map / Video / Video segmentation and tracking
Hierarchical Semantic Contrast for Scene-Aware 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 |
|---|---|---|---|
| Add memory-based contrastive learning to both streams (ablation) | HSC scene-aware autoencoders · Avenue | AUC (%): 91.1 -> 92.4 | Helped |
| Add memory-based contrastive learning to both streams (ablation) | HSC scene-aware autoencoders · ShanghaiTech | AUC (%): 80.7 -> 83.0 | Helped |
| Add motion augmentation (ablation) | HSC with both streams and MemCL · Avenue | AUC (%): 92.4 -> 93.7 | Helped |
| Add motion augmentation (ablation) | HSC with both streams and MemCL · ShanghaiTech | AUC (%): 83.0 -> 83.4 | Helped |
| Combine scene-appearance and scene-motion streams (ablation) | HSC without MemCL or MA · Avenue | AUC (%): 90.6 / 81.3 -> 91.1 | Helped |
| Add scene-aware feature encoding (ablation) | HSC autoencoder without contrastive learning · Avenue | AUC (%): 91.0 -> 90.9 | Hurt |
| Add scene-aware feature encoding (ablation) | HSC autoencoder without contrastive learning · ShanghaiTech | AUC (%): 80.8 -> 80.2 | Hurt |
| Add scene-aware feature encoding (ablation) | HSC autoencoder without contrastive learning · Mixture [01,02] | AUC (%): 78.6 -> 80.5 | Helped |
| Add object-centric decoding (ablation) | HSC autoencoder without contrastive learning · Avenue | AUC (%): 90.9 -> 91.1 | Helped |
| Add scene-level contrastive loss (ablation) | HSC with scene-aware encoder and object-centric decoder · ShanghaiTech | AUC (%): +0.9 | Helped |
| Add object-level contrastive loss (ablation) | HSC with scene-aware encoder and object-centric decoder · ShanghaiTech | AUC (%): +1.3 | Helped |
| Add scene-level and object-level contrastive losses together (ablation) | HSC with scene-aware encoder and object-centric decoder · ShanghaiTech | AUC (%): 80.7 -> 82.2 | Helped |
Closest papers by meaning
- VadCLIP: Adapting Vision-Language Models for Weakly Supervised Video Anomaly Detection (AAAI 2024 · similarity 0.87)
- Open-Vocabulary Video Anomaly Detection (CVPR 2024 · similarity 0.87)
- Harnessing Large Language Models for Training-free Video Anomaly Detection (CVPR 2024 · similarity 0.86)
- Learning Event Completeness for Weakly Supervised Video Anomaly Detection (ICML 2025 · similarity 0.86)
- Video Anomaly Detection with Motion and Appearance Guided Patch Diffusion Model (AAAI 2025 · similarity 0.86)
- VADTree: Explainable Training-Free Video Anomaly Detection via Hierarchical Granularity-Aware Tree (NeurIPS 2025 · 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.