ProDisc-VAD combines prototype-based normality modeling with contrastive learning on extreme-scoring instances, hitting 97.98% and 87.12% frame-level AUC with a 0.4M-parameter head.
Weakly-supervised video anomaly detection with snippet anomalous attention,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ProDisc-VAD: An Efficient System for Weakly-Supervised Anomaly Detection in Video Surveillance Applications
ProDisc-VAD combines prototype-based normality modeling with contrastive learning on extreme-scoring instances, hitting 97.98% and 87.12% frame-level AUC with a 0.4M-parameter head.