Abstract
Application of deep learning face recognition network using CNN VGG-16 convolutional neural network with three self-collected image data sets, VNPT Dong Thap staff and request relevant partners to provide images of group members traveling to BTS stations in the area of VNPT Dong Thap to perform the work as agreed, these image files are used for training and testing face recognition, to decide whether to allow access to the station and like controlling the most scientific work schedule. In this study, the CNN VGG-16 network image and the CNN VGG-19 network are applied for face recognition, analysis and comparison between these two networks to evaluate the performance of face recognition with the highest results. The experimental process of studying the training algorithm is RMSprop, Adam and also need to compare the accuracy with other network models performed on the face dataset VNPT Face for the purpose of improvement. future face object recognition performance.