Crowd Counting Based on CSI and Convolutional Neural Network

With the rapid development of people-centered internet of things technology, crowd counting has attracted more and more attention and plays an important role in smart home and public security management. This paper present a crowd counting system based on WiFi Channel State Information (CSI) and Convolutional Neural Network (CNN). The system uses the convolutional neural network to process input data and achieves a recognition accuracy of 97% in a six people environment. We analyze the influence of distance on recognition accuracy and discuss the data processing methods with the other two neural networks. Furthermore, we compare this system with the other system in the similar parameters and validate the effectiveness of our system in crowd counting using CSI.