This project, developed for Computer Vision and Cognitive Systems course, introduces a self-service checkout system designed for canteens, leveraging deep learning techniques for automatic food recognition. The system aims to optimize self-canteen operations and enhance the overall customer experience by combining object detection and classification networks. The system is capable of handling varying numbers of food items by generating embeddings for detected objects and matching them to a menu database, enabling real-time classification without prior knowledge of the number of food categories.