Food Recognition and Nutrition Estimation using MobileNetV2 CNN architecture and Transfer Learning
Karima Moumane, Ikram El Asri, Taoufiq Cheniguer, Soufiane Elbiki
Abstract
A food detection and nutritional analysis project is a cutting-edge tool that allows users to easily identify the nutritional information of different types of food by simply taking a photo of the food item. This paper leverages state-of-the-art image recognition technology to analyze the image and extract crucial information such as calorie count, protein content, and other relevant nutritional facts. To accomplish this, a Convolutional Neural Network (CNN) model based on the MobileNetV2 architecture has been trained, and it can successfully identify 190 different food categories, including a wide range of cuisines, both Western and local. Transfer Learning has been employed to achieve accurate and robust food detection across an extensive array of food categories. Furthermore, a mobile app has been developed to facilitate food identification and provide a wide range of nutritional information to users based on the use of the Edamam food API.