SmartDine: A Comprehensive System for Enhancing Customer Satisfaction and Optimizing Business Operations in the Restaurant Context

This research explores the integration of real-time video analysis, deep learning techniques, and advanced data processing tools to enhance customer satisfaction and operational efficiency in restaurant management. Leveraging CCTV footage, along with tools such as Python,PyTorch, OpenCV, and advanced models like YOLOv10 for object detection and YOLO-pose for key point extraction, the study focuses on optimizing key aspects such as food satisfaction analysis, customer engagement and wait time reduction, optimal use and planning of resources. Custom models, including an LSTM network, analyze detected key points to identify customer actions and estimate time spent on various activities. Data augmentation techniques enhance model performance, with metrics like precision and recall used for evaluation. The study further aims to provide restaurant managers with a dynamic dashboard that tracks real-time insights, including table turnover rates, preferred menu items, and optimal resource allocation. This research contributes to data-driven restaurant management by offering a robust framework for real-time operational analysis, ultimately improving decision-making and enhancing the dining experience.