DEEP LEARNING BASED FOOD RECOGNITION

Abstract

Literature has indicated that accurate calorie assessment is very important for assessing the effectiveness of food. With the help of mobile devices and rich cloud services, it is now possible to develop new computer-aided food recognition system for calorie assessment. Food recognition is a difficult problem, because food is deformable and exhibits high intra-class variation. However, due to the nature of food images, their recognition is a particularly challenging task, which is why traditional approaches in the field have achieved a low classification accuracy. Deep neural networks have outperformed such solutions. We propose a deep convolutional neural network that performs food image classification and calorie detection on the base of training dataset with their labels. With the help of this CNN model that we built, we have got the training accuracy of 51% with 40 epochs but we have got the testing accuracy depends on the image that is predicted by the model which is taken on real-time. We are taking image of 10 classes offood and respectively testing our model on real-time.