Implementation of a Smart Checkout System Based on Face Recognition Using YOLO v3-tiny
Chin-Fa Hsieh, Pin-Hong Liu, Jai-Xuan Lai, Shih-Hao Sun · 2022 IEEE 11th Global Conference on Consumer Electronics (GCCE) · 2022
With the application of biometrics such as voice, fingerprint, face, and iris recognition, electronic systems are becoming more and more automated and intelligent. The reduction of paper use accordingly results in the effect of environmental protection, by protecting the forest resources as well as reducing the environmental pollution. Based on these two goals, face recognition technology and Line communication software, being inseparable in life, are combined in this paper to complete an intelligent checkout system. A face recognition system based on YOLOv3-tiny is then built in this paper. In this paper for face recognition, the WIDER FACE database is used for training and the GoogLeNet is used for improving the recognition rate. The result shows that the accuracy rate for this improved model is 7.57% higher than without the GoogLeNet one. The combination of recognition result and LINE Notify function can then automatically transmit the consumption details, at a mode of paperless environmentally friendly shopping checkout.