MealMind: An AI Framework for Reducing Household Food Waste and Decision Fatigue Through Automated Inventory Management

This paper introduces MealMind, an integrated AI framework designed to address two interconnected challenges in household management: Decision Fatigue in meal planning and significant financial losses from preventable food waste. The system implements a three-component architecture: (1) Computer Vision (CV) for auto mated fridge inventory via smartphone scanning, eliminating the 59.3% failure rate of manual entry; (2) a Stock Optimization Algorithm using a Spoilage Proximity Index (S) to prioritize soon-to-expire items; and (3) a Hybrid LLM Planner gen erating personalized meal plans from available inventory, including comprehensive 7-day menus and special event planning for guests. Our mixed-methods study (survey: N=82 complete responses; interviews: N=5) quantifies the problem: 57.3% of households face daily "what to cook?" stress, 52.4% discard food due to forgetfulness, 78% of waste comprises expired items, and 58.5% demand better tools for weekly meal planning. We present a functional mobile prototype demonstrating technical feasibility and propose two testable hypotheses: MealMind reduces weekly planning time by >40% (H1) and decreases financial waste by >30% (H2). The paper concludes with a rigorous experimental design for val idation, positioning MealMind as a foundational layer for sustainable, intelligent kitchen ecosystems.