What question did this study set out to answer?

The aim is to develop a data-driven framework for detecting stampedes and classifying crowd risks using advanced machine learning techniques.

April 12, 2026Scientific ReportsOpen Access

Stampede detection and crowd analysis using CNN-LSTM and farneback optical flow

View Full Paper

Ask AI

Bookmark

The aim is to develop a data-driven framework for detecting stampedes and classifying crowd risks using advanced machine learning techniques.
Utilized a hybrid CNN-LSTM architecture for processing crowd data.
Integrated Farneback optical flow for enhanced motion analysis.
Combined UCSD Anomaly Detection and Agoraset datasets for diverse crowd behavior modeling.
Trained on 10,000 annotated frames with augmentation for robustness.
Achieved an accuracy of 99.75% in stampede detection and risk classification.
Classified crowd states into four levels: normal, moderate, dense, and risky.
Demonstrated robustness during cross-dataset evaluations on the UMN benchmark.

No takes yet. Share an insight, caveat, or question.

Demonstrates an automated stampede detection framework in crowded environments, suggesting improved public safety measures.

Bhola et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark