Virtual Booth

NutriVision is an AI-powered system designed to estimate nutritional information of Malaysia food from a single image. Addressing the lack of standardized local food datasets and limitations of existing weight estimation methods, NutriVision integrates object detection, ArUco marker-based measurement, and XGBoost regression model to predict food weight accurately. Next, nutritional data is retrieved from a self-designed Malaysia food composition database API. The system has achieved a low Mean Absolute Error (MAE) of 6.67g and Mean Absolute Percentage Error (MAPE) of 5.28%, significantly improving upon existing solutions. NutriVision supports healthcare and personal use, offering a scalable and practical solution for medical applications.