A review of vehicle annotation method for urban traffic surveillance

At the beginning of its presence, the Intelligent Transportation System (ITS) played a significant role in assisting vehicle surveillance because of its ability to recognize a vehicle license plate. However, in conditions of dense urban traffic, license plate information can be challenging to obtain. Since license plate information probably cannot be obtained, the vehicle's fine-grained features such as color, model, year, and manufacturer, are considered necessary. However, the effort to obtain fine-grained feature information manually is exhaustive. Some researchers have begun to develop a method to obtain these attributes more efficiently, namely vehicle annotation method. This paper provides a review of the vehicle annotation method and the preprocessing method and dataset characteristics commonly used in vehicle annotation method research. Besides highlighting some of the methods compared, we will also discuss issues found in current research and future research to address existing issues at the end of this paper. © 2021 ICIC International. All rights reserved.