Face age and gender recognition based on improved VGGNet algorithm

Abstract

School of Mathematics and Statistics, Nanjing University of Information Science & Technology, \u00a0 Nanjing, 210044, China (Received February 20 2019, accepted June 24 2019) Recognition \u00a0of \u00a0age \u00a0and \u00a0gender \u00a0based \u00a0on \u00a0face \u00a0image \u00a0is \u00a0one \u00a0of \u00a0the \u00a0hotspots \u00a0of \u00a0current \u00a0artificial intelligence \u00a0research. \u00a0In \u00a0this \u00a0paper, \u00a0an \u00a0improved \u00a0VGG+SENet \u00a0algorithm \u00a0is \u00a0proposed \u00a0to \u00a0simplify \u00a0the identification of age \u00a0and gender \u00a0algorithm by simplifying \u00a0VGGNet model, improving the loss \u00a0function \u00a0and embedding the SENet module. Compared with other models, the improved network structure and loss function model proposed in this paper can quickly and accurately obtain output recognition results. Experimental results on \u00a0multiple \u00a0benchmark \u00a0face \u00a0datasets \u00a0show \u00a0that \u00a0the \u00a0proposed \u00a0improved \u00a0VGG+SENet \u00a0algorithm \u00a0has \u00a0higher recognition accuracy than other related models based on deep learning.