Abstract:
Working mothers entrust their children to daycare centers, which in turn provide parents with a detailed report of the food given to the child throughout the day. This research work develops an application that uses the image of a nutrient to calculate its caloric value. It assists the parents to make sure that their child is consuming the right amount of nutrition. Taking the right amount of nutrition is necessary for the development of a child, as every nutrition plays a significant role in growth. For this purpose a Meal Recommendation System is designed, This system takes the uploaded image by the user as an input and recognizes the food. For the food recognition, different models like CNN, Inception V3, MobileNet, yolov5, and yolov8 are trained on FOODD, ECUSTED, and Food-360 datasets. From the findings of the research work it is revealed that yolov8 comparatively gives better results than other models. The area of the image is calculated using image segmentation and then micro-nutrient values of the recognized food are calculated from openfoodfacts dataset. These recognized foods along with mass and micro-nutrient values are saved in the food log. From the food log, we calculate the total intake of the child, and from the age group we find the required nutrient values given by dietician and by subtracting total intake from requirements deficiencies are calculated. Consequently, The deficiencies are calculated and data from the openfoodfacts helps to recommend the foods that are high in those nutrient values in deficiencies. Hence, the appropriate food is recommended. The recommended food is validated by looking at the values which they highly contained.
