Video anomaly detection suffers from motion observation quality shortcuts
Auditing Observation-Quality Shortcuts in the Evaluation of Person-Centered Motion for Video Anomaly Detection
Computer Vision and Pattern Recognition
Summary
Detecting unusual events involving people in videos is important but tricky. The authors show that current methods often rely on how well people are seen rather than real unusual movements. They develop a way to check if these methods are truly detecting abnormal motion or just benefiting from video quality. Their tests reveal that high detection scores can be due to differences in person detection and tracking quality, not actual motion differences. This means better tests are needed to trust that models understand unusual human motion.
What this means in practice
- •For video surveillance teams: Improve anomaly detection systems by auditing and controlling for person observation quality to avoid false alarms caused by video quality rather than actual abnormal events.
- •For security software engineers: Enhance video monitoring tools by integrating explicit checks on person detection and tracking reliability to ensure anomalies reflect real motion changes.
Authors
Chaehyun Kim, Changwon Lim
Abstract
Video anomaly detection (VAD) identifies deviations from normal patterns, but such deviations do not necessarily correspond to safety-critical events. We focus on human-centered emergencies and violence and investigate whether motion-based anomaly discrimination reflects task-relevant motion information or observation-quality differences arising from person detection, pose estimation, and tracking. We introduce an observation-quality-aware evaluation framework that combines quality diagnostics, quality balancing, motion orthogonalization, and robustness analyses. On human-centered categories of UCF-Crime, quality-only features match or exceed the discrimination of multi-cue motion in the original cohort, revealing a strong observation-quality shortcut. After cardinality matching brings all 11 measured quality variables below a prespecified marginal balance threshold, multi-cue motion consistently outperforms quality-only features across Logistic Regression, Random Forest, and RBF-SVM, yielding ROC-AUC gains of 0.046-0.141. Because nonlinear quality-only prediction remains above chance after matching, we interpret this result as additional discrimination beyond measured marginal quality rather than as evidence of quality-free or causal motion understanding. These findings show that high VAD performance should not be interpreted as evidence of abnormal-motion modeling without auditing the reliability of the person-observation pipeline. Our study provides a systematic protocol for evaluating observation-quality shortcuts and for assessing whether motion-based discrimination survives explicit control of detection, pose-estimation, and tracking reliability.