Towards a home Digital Nutritionist: a proof of concept for AI-guided adaptation of children’s meals

This thesis presents a digital nutrition system that adapts home-cooked recipes to the dietary needs of children aged 3 to 9 years. Using the official nutritional guidelines of the Veneto Region, the system automatically extracts ingredients, adjusts portion sizes, and ensures nutritionally appropriate meals. Developed in Python, it integrates rule-based logic, machine learning, and generative AI, supported by a local SQL database. Results show that the system reliably produces safe and guideline-compliant recipe adaptations, offering families a transparent and evidence-based tool for everyday meal planning.