Food Detection and Nutritional Recognition System Using Neural Networks

Abstract

The purpose of this research paper is to implement a user preference-based recommendation system that not only detects food items but also gives nutritional content of the food specially fruits, diet suggestions for the targeted calorie values, and special diet for diseased persons and also the combination of food that harms the digestive system. The proposed methodology incorporates a model for the food varieties specially fruits using deep learning and convolutional neural networks (CNN). The neural network model takes input as image and text data, analyzes it by using SoftMax activation function it provide multi classification and gives the nutritional values. The methodology uses stochastic gradient descent (SGD)which is a simplified optimization algorithm for large-scale datasets. The output values are displayed through a dedicated website designed to show the nutritional contents, user recommendation diet plan, disease-based diet plan, etc. In addition to this, the proposed paper focuses to help the people to improve their dietary habits and lead to the minimal health risks.