A YOLO-based Method for Improper Behavior Predictions

Deep learning (DL) algorithms have found extensive applications in the realm of intelligent policy contexts. In specific environments like waiting rooms and interrogation rooms, where actions such as physical assault and confrontations can pose serious risks, identifying and addressing improper behaviors is crucial. YOLO stands out as a prominent network architecture commonly employed for object detection tasks. In addition, SLTM-GCN represents an innovative DL model designed for predictive analysis utilizing unstructured data. To identify instances of inappropriate behavior in videos, this research leveraged both the YOLO and SLTM-GCN DL models. Consequently, this study makes a significant contribution by establishing the relevance of DL in the context of smart policing.