Customer flow and behavior analysis using YOLO, DeepSORT, and computational geometry algorithms

In today’s competitive market, understanding customer flow and behavior is crucial for optimizing marketing strategies, ultimately leading to increased business revenue. We present an automated system for custom flow and behavior analysis utilizing YOLOv8, DeepSORT and computational geometry algorithms, with minimum user intervention. Using surveillance video footage from real-world scenarios, such as department stores, our system can generate analysis reports that yield data visualization and provide valuable insights into customer flow and behavior, thereby offering informed guidance for store managers to optimize marketing strategies. Technical approaches include YOLOv8 for object detection, DeepSORT for object tracking, and computational geometry algorithms for analyzing customer flow and behavior. Our research demonstrates promising results by integrating video processing and AI techniques in retail analytics, thereby providing valuable insights to enhance the shopping experience and drive revenue growth.