A Review on Real-Time Person Detection and Description Systems

Advancements in surveillance technologies are pivotal in addressing the growing demands for real-time monitoring and actionable insights across industries such as security, public safety, and traffic management. This review explores existing research and technologies in real-time person detection, attribute analysis, and semantic description generation, highlighting their applications, limitations, and gaps. It further presents a novel system integrating YOLOv8 for object detection, OpenCV forattribute analysis, and Qwen2VL for vision-language tasks, addressing key challenges such as latency, scalability, and dynamic environment adaptability. Future directions include integrating emotion recognition, multilingual GUIs, and predictive analytics to enhance system utility and versatility.