Malaysian food recognition and calorie counter application

A calorie counter application is being widely used by people around the world. However, the options for Malaysian food items are limited in those applications. This paper explores the idea to add auto recognition feature into a calorie counter application while identifying the challenges faced to perform recognition on Malaysian traditional food. The prototype described in this paper is able to recognize five different types of Malaysian `kueh' or dessert. The main method of feature extraction is by generating the color histogram and an artificial neural network is used to classify the food.