Vision-Based Consumer Behavior Analysis System for Retail Optimization

Retailers are businesses that sell products and services to consumers for personal or family use, acting as the final step in the consumer supply chain. Unlike wholesale transactions, retail sales are motivated by personal or family satisfaction derived from the final consumption of purchased goods. In today's competitive market, retail managers must understand their operational environment to develop effective strategies. Understanding why consumers choose to purchase, how they select stores, and how they choose products is essential for success. This research focuses on the development of a system for analyzing consumer purchasing behavior to gain deeper insights into consumer actions and enhance retail operations. The proposed system is designed to detect customer behavior using video data, either in real-time or from recorded footage. It allows for flexible detection zones, heatmap generation, and classification by gender and age. The system utilizes YOLOv8 for object detection and integrates DeepFace for image processing, achieving a high accuracy rate of 90.7%. The findings of this research are expected to support retail businesses in improving their store layouts and marketing strategies more effectively.