A SURVEY OF ARTIFICIAL INTELLIGENCE TECHNIQUES FOR PERSONALIZED NUTRITION AND DIETARY ASSESSMENT

Abstract

The increasing prevalence of obesity, diabetes, cardiovascular diseases, and other nutrition-related disorders has highlighted the need for personalized dietary interventions that consider individual health characteristics, lifestyle behaviors, and nutritional requirements. Traditional dietary assessment methods often depend on self-reported food records and generalized dietary guidelines, which may not adequately capture individual variability in dietary habits and metabolic responses. In recent years, Artificial Intelligence (AI) has emerged as a promising technology for enhancing personalized nutrition and dietary assessment through intelligent data analysis, predictive modeling, and automated decision-making. AI techniques such as Machine Learning (ML), Deep Learning (DL), Computer Vision, Natural Language Processing (NLP), and Recommender Systems have demonstrated significant potential in analyzing dietary patterns, estimating nutrient intake, recognizing food items from images, predicting health outcomes, and generating individualized nutrition recommendations. Furthermore, the integration of wearable devices, mobile health applications, electronic health records, and omics data has enabled the development of comprehensive nutrition monitoring systems that support real-time dietary management. This survey presents a detailed review of AI techniques applied to personalized nutrition and dietary assessment, emphasizing their methodologies, applications, datasets, advantages, and limitations. The study also examines current challenges, including data privacy, model interpretability, dataset heterogeneity, and clinical validation. Additionally, emerging trends such as explainable AI, federated learning, multimodal learning, and AI-driven precision nutrition are discussed. The findings indicate that AI-based nutritional systems have the potential to transform dietary assessment and personalized healthcare by providing accurate, scalable, and data-driven nutritional guidance. However, further research is required to improve model transparency, reliability, and real-world applicability before widespread adoption can be achieved.