PRW, short for Person Re-identification in the Wild, is a large-scale dataset for end-to-end pedestrian detection and recognition in outdoor videos. It contains 932 pedestrians and 11,816 video frames, where pedestrians are annotated with bounding box locations and identities. The dataset uses 6 synchronized cameras to collect videos and uses PRW technology to evaluate outdoor person re-identification.

This dataset can be used to develop sophisticated multi-query or re-ranking methods.