Food Image Recognition based on ResNet

Abstract

Image classification has always been one of the basic tasks in the community, which has been widely applied in many fields, such as the food recognition. As the key technology of dining robot, the food image recognition aims to predict the category of food in the given image, which has attracted a lots of research attentions from both the academia and industry. Early efforts of food image recognition mainly rely on the manual features, whose accuracy cannot meet practical application requirements. Thanks to the rapid development of convolutional neural networks, food image recognition based on deep learning has made breakthroughs in both accuracy and speed. In this paper, we propose a food image recognition method based on the ResNet. Extensive experiments demonstrate the effectiveness of our method, which can provide some new insights for the automatic food recognition.