An AR-Assisted Food Ingredient Recognition System for Smart Refrigerators Using Deep Learning

Food waste at the household level has become a significant global issue, largely due to inefficient food management and lack of awareness of stored ingredients. This paper proposes an Augmented Reality (AR)-based food recognition system designed for household refrigerators. The system integrates lightweight deep learning-based food recognition with real-time AR visualization to improve food visibility and usage efficiency. Using a lightweight CNN-based object detection model (YOLOv8) with real-time AR visualization, the proposed system identifies common refrigerator food items under challenging conditions such as low lighting, occlusion, and clutter. Experimental results demonstrate that the system can accurately recognize and localize multiple food items in real time, providing a practical solution for reducing household food waste and enhancing smart kitchen applications.