AN INTELLIGENT FOOD ADVISORY SYSTEM USING ARTIFICIAL INTELLIGENCE

Abstract

The growing global concern for healthier living has drawn significant attention to health and wellness worldwide. The World Health Organization(WHO) has acknowledged the rise in noncommunicable diseases (NCDs) like premature heart disease, cancer, and diabetes, with unhealthy diets being a major contributing factor. In response to this challenge, there is a strong global need for simple, intelligent tools that assist individuals in making informed dietary choices and understanding the nutritional impact of their meals. This research proposes the development of an Intelligent Food Advisor, an integrated webbased system designed to bridge the gap between meal consumption and nutritional awareness. The system leverages Artificial Intelligence (AI) to provide instant food recognition, detailed nutritional analysis, and customized dietary recommendations by combining real-time data processing with sophisticated deep learning models. The Intelligent Food Advisor empowers individuals to make informed dietary choices, promotes healthier eating habits, and contributes to addressing the worldwide issue of dietrelated diseases. The system architecture combines a frontend built with HTML, CSS, and JavaScript, with a robust backend powered by the FastAPI framework and an SQL relational database for user data management. At its core, the system employs an EfficientNetV2B2 deep learning model finetuned on the Food101 dataset to accurately identify food items from useruploaded images. Upon recognition, the system utilizes the Mistral 7B Large Language Model (LLM) via an API to generate realtime nutritional information and actionable advice. Additionally, the application provides users with graphical charts that categorize food choices as healthy, moderately healthy, or unhealthy. This comprehensive workflow from user authentication to daily dietary logging and performance monitoring makes nutritional management seamless, intuitive, and engaging.