NutriEar: Robust Nutrition-Aware Food Classification from In-Ear Acoustic Signals

Abstract

Convenient tracking of food intake is essential for linking diet to health, enabling personalised nutrition guidance, early metabolic risk detection, and effective prevention of chronic disease. Recent wearable sensing advances have begun to automate eating monitoring. However, these systems largely focus on detecting when users eat and only weakly address what they eat. In particular, state-ofthe-art solutions typically cover a narrow set of foods or textures and rely on strong assumptions about individual eating behaviour. Moreover, they overlook the underlying nutritional implications, the information most relevant to end users, thereby limiting the usefulness of their outputs for real-world dietary guidance. In this paper, we present NutriEar, an in-ear audio sensing system that performs nutrition-aware classification of food intake from chewing sounds. Rather than recognising arbitrary food types, NutriEar maps in-ear acoustics to an eight-class nutrition-texture taxonomy grounded in food science, jointly capturing dominant macronutrient role and mechanical texture. NutriEar records in-ear audio during eating, applies lightweight segmentation of chewing events, and derives a hybrid representation that combines engineered acoustic features with learned embeddings from a supervised contrastive learning framework, enabling a compact nutrition-aware food classification pipeline. An evaluation on a dataset we collected with over 30 food types from 15 users under varied eating conditions demonstrates that NutriEar achieves 80.18% average leave-one-subject-out (LOSO) accuracy and outperforms state-of-the-art baselines. These results highlight the untapped potential of earable audio sensingas a practical pathway toward everyday dietary monitoring with meaningful nutritional insights.