Calories analysis of food intake using image recognition

Abstract

In recent year, healthy is a topic that people concern. It is obviously that eating the food with high amount of calories cause several problems to our health. Recording the amount of calories of the food intake in each meal is one of the stretchy to solve such problem. Although the people can record their meal and discuss with doctors or experts, it is not so convenient and they cannot know the amount of calories before the meal. This paper presents a technique of image processing to recognize images of food taken by users. From the input food images, the users can understand the amount of calories they will take in each meal by using the proposed algorithm. Our method creates feature vector using several features about texture and color, then classify the food images using SVM. In this study, we focused on Thai food. To train the SVM, we group the example food images by food type and the amount of calories. We conduct the experiment to evaluate the performance of the proposed algorithm for both groups of example food.