A novel approach for diet recommending system using faster RCNN and tensorflow lite.

Obesity is a global health issue which is growing rapidly. This results in an unhealthy diet containing high amounts of sugar, salt and fats in addition to an inactive lifestyle. The problems of chronic diseases like obesity and overweight are caused fundamentally because of the vitality of lopsidedness between calories ingested and calories burned. In this paper, an approach is proposed that incorporates the possible mitigation solutions to detect food images and suggest the calorie content using Faster RCNN and TensorFlow Lite for users. Faster RCNN model is first trained using multiple food images and then converted to TensorFlow Lite model which is compatible with Android devices and requires less computational and cloud resources for detecting objects. Real time food and calorie detection along with appropriate meal plan recommendations will definitely change the day-to-day performance of an individual. [ABSTRACT FROM AUTHOR]