.jpg)
LogMeal NuraHealth
We developed and validated a computer-vision pipeline that converts meal images into standardized, machine-readable food data—detecting items on a plate, identifying foods at dish/ingredient level, and producing structured outputs (labels, portions/serving estimates, and nutrition signals) suitable for real-time monitoring. The key scientific result is robust recognition in uncontrolled, real-world conditions (variable lighting, occlusions, mixed dishes) with consistent performance at scale, enabling objective measurement of intake and food-service outcomes without manual annotation.
Practical applications of the prototype/product include:
Hospital and clinical nutrition monitoring
Capture meals at serving time to quantify intake, track adherence to dietary plans, and support care pathways for chronic conditions.
Sports performance and athlete fueling management
Provide objective meal logging and macro/micronutrient estimates to help coaches and nutritionists optimize fueling, recovery, and compliance.
Corporate canteens and contract catering optimization
Measure what is served and consumed to improve menu planning, reduce waste, and standardize reporting across sites.
Public-sector food services (schools, hospitals, residences, prisons)
Generate auditable evidence for quality control, procurement compliance, and sustainability KPIs with minimal operational burden.
Food waste analytics and continuous improvement programs
Establish baselines, identify waste drivers (portioning, unpopular dishes, timing), and track the impact of corrective actions over time.
Integration layer for digital food ecosystems
Feed structured food and nutrition data into existing systems (POS, menu management, BI dashboards, patient records, or wellness apps) via API for alerts, reporting, and decision support.
