Application of CNN-based Face Recognition Technology in Smart Logistics System

Abstract

Smart logistics combines the Internet of Things with big data, cloud computing and other information processing technologies to improve all aspects of logistics business and optimize the status quo of logistics services. This research aims to study the application of Convolutional Neural Network-based face recognition technology in intelligent logistics systems. First, briefly describe the key concepts of Faster Region with CNN feature, convolutional neural networks, the essence of smart logistics, and the background of the overall architecture design of CNN-based smart face recognition logistics, including data modules, data storage modules, and data presentation layers. Analysis of bottlenecks in other smart logistics systems, such as time-consuming large-capacity storage. By choosing WIDER Face as the sample training set and FDDB as the sample test set, the traditional face detection algorithms are compared. Experimental results show that CNN has the fastest detection speed. And the loss rate of CNN is low, and the accuracy can reach 91% under the condition of unlimited faces.