A Cloud Edge Collaboration of Food Recognition Using Deep Neural Networks

Abstract

Deep neural network-based learning methods are commonly used for classifying images or object detection with excellent performances. In this paper, we looked at how effective a deep convolution neural network (DCNN) is to identify food photography. Food identification is a sort of visual fine-grain recognition that is more difficult than traditional image recognition. Mobile apps in many countries have been omnipresent in many aspects of people's lives in the last few years. Fitting the healthcare potential of this pattern has become a focal point in the industry and researchers’ basic applications for architecture that patients should use in their well-being, prevention or treatment method. Mobile cloud computing has been introduced as a possible mobile well-being paradigm interoperability problems management service in various information formats. In this paper, I am integrating deep neural networks with cloud architecture to avoid substantial memory loss in mobile devices or web platforms where I can upload images, and it will predict the actual images and the names of food categories. The datasets I will be using are UECFOOD101; I will be running my deployment on the system as well as on the cloud for retrieving the data easily on mobile phones and web pages. The cloud architecture helps me offload the data, which is not required using computational offloading and profiling.