MoRe: A Large-Scale Motorcycle Re-Identification Dataset

Motorcycles are abundant objects in the transit and often related with traffic and criminal law breaking in many countries. For instance, motorcycles are used to perform rob-and-run crimes and related to an increasing number of accidents that result in severe injuries and deaths. Therefore, it is crucial to identify the motorcycles and monitor its behavior in wide areas. Since off-the-shelf license plate recognition systems only work in specific camera-views, it is important to develop a motorcycle re-identification system based on appearance features.

To supply this need, we propose a large-scale dataset for Motorcycle ReID, MoRe, composed by real surveillance images.

Data Data

Motorcycle Re-Identification (MoRe) dataset, is the first large-scale motorcycle ReID database captured by urban traffic cameras. Precisely, MoRe contains 3,827 distinct identities and 3,478 distractors captured by ten surveillance cameras in a total of 17,619 detected bounding-box images. The dataset presents important challenges to the research community due to the drastic changes of appearance that motorcycles endure between surveillance cameras as a result of different camera resolutions and views, aligned with the distinct illumination conditions, possible occlusions and the inter-class similarities

Data Request:
To be able to download the dataset, please read carefully this agreement, fill it and send it back to one of the suggested e-mails.

All codes used for evaluating ReID task on MoRe data can be found here.