Classification of Turkish Cuisine With Deep Learning on Mobile Platform

In recent years, image classification has been a subject that is often dealt with in computerized vision. This is mainly due to the fact that more accurate results can be achieved in less time by using the classification method. In this study, food classification process has been carried out by using Convolutional Neural Network (CNN) that is frequently preffered deep learning techniques in machine learning issues. According to the obtained results, the highest accuracy rate was obtained by using Stochastic Gradient Descent (SGD) optimization algorithm with Softmax activation function. That accuracy rate was calculated as 93%. A conversion is applied to trained model in order to use in mobile devices. Application was used to recognize food whose photo taken and/or selected. In addition, for food name having the highest recognition percentage; calorie, nutritional values, recipe ingredients and recipe were given. Our study ‘Türk Mutfağı’ was made ready to applying food classification in iOS devices and usage in daily life.