Global Context Assisted Structure-Aware Vehicle Retrieval

In vehicle retrieval, the vehicle patch should first be localized to remove the irrelevant background information. Moreover, the negative samples are much more prevalent than the positive samples, and the information from the negative samples is not fully exploited in the triple loss. What we need is a way to incorporate global knowledge and structure information to address these two issues. Therefore, we introduce a local-global context network for landmark alignment to update the predicted results by using the semantic information and the local compatibility and propose a structure-aware quadruple loss to use multiple and diverse negative samples in retrieval. Experiments on the VehicleID and the ENJOYOR vehicle retrieval datasets demonstrate that our approach obtains accuracy comparable to state-of-the-art approaches in vehicle retrieval.

I. Introduction

Vehicle retrieval is an important topic for both academia and industry, and it has been an active area of research over the past decade. In this paper, we focus on the law enforcement application. Since a suspect may change a license plate to avoid being found, an effective vehicle retrieval mechanism based on the content of an image can help the police to locate a missing vehicle and catch the suspect, even if the license plate is changed by the suspect.