This work addresses the problem of vehicle re-identification through a network of non-overlapping cameras. As our main contribution, we propose a novel two-stream convolutional neural network (CNN) that simultaneously uses two of the most distinctive and persistent features available: the vehicle appearance and its license plate. This is an attempt to tackle a major problem, false alarms caused by vehicles with similar design or by very close license plate identifiers. In the first network stream, shape similarities are identified by a Siamese CNN that uses a pair of low-resolution vehicle patches recorded by two different cameras. In the second stream, we use a CNN for optical character recognition (OCR) to extract textual information, confidence scores, and string similarities from a pair of high-resolution license plate patches. Then, features from both streams are merged by a sequence of fully connected layers for decision. As part of this work, we created an important dataset for vehicle re-identification with more than three hours of videos spanning almost 3,000 vehicles. In our experiments, we achieved a precision, recall and F -score values of 99.6%, 99.2% and 99.4%, respectively. As another contribution, we discuss and compare three alternative architectures that explore the same features but using additional streams and temporal information. The proposed architectures, trained models, and dataset are publicly available at https://github.com/icarofua/vehicle-ReId .
View on arXiv