Sustainable Smart Advertisement Display Using Deep Age and Gender Recognition

Abstract

We propose an innovative and sustainable advertisement display system tailored to bystanders' liking based on age and gender using Convolutional Neural Networks. The bystander's face is detected using a webcam attached to the advertising screen and fed into the trained Convolutional Neural Network to identify the age and gender. The classification results are stored in a real-time database server and retrieved on a website that enables the client to customize advertisements to the targeted groups. The website provides an easy-to-use interface to control the system contents and report daily statistics. This loop, starting from detecting bystanders' faces until the targeting advertisements are displayed, takes place daily every two seconds. Moreover, a motion sensor is connected to the advertising screen for sleep mode to ensure the system's sustainability. The proposed system serves as an intelligent advertisement platform that guarantees the proper content delivery to the right people. The proposed system achieved accuracy is of 91.7% and 95.5% for age and gender, respectively. This developed system could also display pieces of advice and news in populated public places, which will increase awareness and knowledge in society.