CalorieSense: Design and Development of a Barcode-Free Calorie Monitoring Device using MobileNetSSD for Automated Food Detection

In a fast-paced environment, where unhealthy eating habits become a norm, monitoring calorie consumption and maintaining a nutritious diet is vital. Engaging in physical activities, preparing well-balanced meals, and monitoring calorie intake are crucial to handle these concerns. These practices impact overall health and prevent chronic disease risk to maintain a healthy body-mass index (BMI). Nevertheless, accurately controlling the calorie content of unpackaged food products poses challenges. Previous research proposed barcode-based solutions. However, such solutions were limited to barcode-based products with a wide range of food unrecognized. We introduced CalorieSense, an innovative prototype calorie monitoring device. CalorieSense comprised two key elements. The first one was object detection through the MobileNetSSD model in combination with a load cell sensor to estimate the calorie content of selected food items. The second component was a mobile application designed to scan the QR code and display detailed information about the food item to facilitate user-friendly calorie monitoring by allowing users. The results demonstrated a feasible approach to detecting food products and estimating their calorie content, achieving an accuracy of 94.55% in detecting test subjects. Furthermore, CalorieSense showed exceptional calorie estimation accuracy, with a minimal 0.94% error from the actual calorie values.