Improved YOLOv4 for Enhanced Public Safety Management amid Civil Unrest

Pedestrian detection is crucial for various applications, including intelligent surveillance, urban management, and public safety. This paper proposes an enhanced YOLOv4 model for improved public safety management in the midst of civil unrest. The model integrates advanced features such as CSPDarknet53 backbone network, SPP module, and optimized loss functions to achieve superior detection performance. We conducted experiments using video data collected from YouTube and evaluated the performance of the model using ROC curves and AUC values. The results show that our proposed method outperforms traditional CNN and SVM methods with significantly higher AUC values, indicating its effectiveness in detecting anomalous behavior. This research contributes to the advancement of pedestrian detection technology and provides valuable insights for future research in public safety management.