November 28, 2025Scientific ReportsOpen Access

Robust crowd anomaly detection via hybrid ensemble learning for real-world surveillance

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Hybrid ensemble learning improves real-time crowd anomaly detection, enhancing robustness and accuracy using deep learning methods.

Achieving near-perfect accuracy of 99.89% in crowd anomaly detection, optimizing ensemble methods with real-time applications.
Utilized optical flow and dimensionality reduction techniques to classify behaviors as normal or abnormal effectively.
Employing Random Forests and Gradient Boosting ensures a robust framework for surveillance in diverse environments.
Demonstrates strong generalizability across various datasets, indicating significant potential for practical surveillance applications.

Mabrouk et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark