| Abstract |
The rising concerns about public safety in urban areas highlight the need for sophisticated surveillance systems aimed at enhancing the well-being and security of all citizens. This paper presents Citizen Safety Analytics, a comprehensive, real-time threat detection solution designed to foster secure public environments through advanced analytics and proactive incident prevention. This system is designed to enhance public security by identifying individuals, assessing emotional cues, and recognizing significant gestures in real-time. By identifying abnormal patterns - such as an individual in a secluded area at night or someone surrounded by a group exhibiting unusual behavior - the platform addresses potential safety risks. By analyzing real-world behavioral patterns, the system can identify unusual movements, group formations, or hostile actions, prompting a quick response. Weapon detection capabilities bolster public safety by identifying physical threats, while audio-based distress detection recognizes calls for help or distress signals. Gesture recognition further enhances responsiveness by detecting SOS signals and enabling prompt action. To facilitate swift responses, the system generates real-time alerts and notifications, such as SMS or Telegram messages, to law enforcement and emergency contacts. Additionally, the platform aggregates data to identify safety hotspots, aiding in strategic urban planning and public security initiatives. Key features include person detection with gender classification, real-time demographic distribution analysis, and anomaly recognition. This proactive analytical approach, supported by real-time alerts and comprehensive detection mechanisms, has the potential to significantly reduce risks and promote public safety for all. © 2025 IEEE. |
