22 datasets found

The SYSU-30k dataset is a large-scale weakly supervised person ReID dataset with over 29 million images gathered from TV program videos. The videos are randomly broken into...

The MUDD dataset was curated from PerformancePhoto.co and manually labeled in the traditional, fully supervised, ReID setting. We adapt the MUDD dataset to the weakly supervised...

The acquisition of large-scale, precisely labeled datasets for person re-identification (ReID) poses a significant challenge. Weakly supervised ReID has begun to address this...

MARS is an extension of the Market-1501 dataset [51]. It has been collected from six near-synchronized cameras. It consists of 1,261 different pedestrians, who are captured by...

The SYSU-MM01 dataset is a popular dataset for visible-infrared person re-identification. It contains 491 identities with 4 VIS images and 2 IR images captured by 4 cameras.

The RegDB dataset is a popular dataset for visible-infrared person re-identification. It contains 412 identities with 10 VIS images and 10 IR images captured by a pair of...

The LLCM dataset is a challenging low-light cross-modality dataset for visible-infrared person re-identification. It contains 46,767 bounding boxes of 1,064 identities captured...

The Duke dataset is a large-scale person re-identification dataset, containing 702 identities with 16522 images.

Person Re-Identification (ReID) aims to recognize a person-of-interest across different places and times. RF-ReID uses radio frequency (RF) signals for longterm person ReID.

Person Re-Identification (ReID) aims to recognize a person-of-interest across different places and times. RF-ReID uses radio frequency (RF) signals for longterm person ReID.

The CUHK03-np dataset comprises 14,097 photos of 1,467 individual identities.

The Market dataset has 1,501 identities captured by six cameras.

Re-identification (Re-ID) problems aim to determine corresponding targets across multiple cameras or different locations in the same camera from the gallery set with a given probe.

The dataset is used for person re-identification and image retrieval.

A dataset for unsupervised person re-identification using Generative Adversarial Networks (GANs).

The RegDB dataset contains 412 identities with 10 visible images and 10 infrared images from a pair of overlapping visible and infrared cameras.

The SYSU-MM01 dataset contains 491 identities captured by 4 visible cameras and 2 infrared cameras both including indoor and outdoor environments.

The Market-1501 dataset is a large-scale benchmark for person re-identification.

Text-based Person Search (TPS) task is targeted on retrieving pedestrians to match text descriptions instead of query images.

Text-based Person Search (TPS) task is targeted on retrieving pedestrians to match text descriptions instead of query images.