Real-Time Vehicle Classification Using MobileNet

Authors: Reagan L. Galvez, Melvin K. Cabatuan, and Argel A. Bandala

Abstract

classification is an important part of vision systems and has several applications like autonomous cars and surveillance. This is a challenging task because computers see images differently from humans. This paper used the MobileNet model for training the data and tested it on an Android device. This model is lightweight and efficient compared with previous developed models. This was inspired by the sample code from Google Codelabs. Experiment results show that the Android application can accurately classify the type of vehicle in real time.