SYSU-30k

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 clips, and then each clip is manually annotated with an identity, but all detected people are noisily assigned that identity, forming bag-level labels.

Data and Resources

Cite this as

Jacob Tyo, Zachary C. Lipton (2024). Dataset: SYSU-30k. https://doi.org/10.57702/e6qt4umf

DOI retrieved: December 17, 2024

Additional Info

Field Value
Created December 17, 2024
Last update December 17, 2024
Defined In https://doi.org/10.48550/arXiv.2402.07685
Author Jacob Tyo
More Authors
Zachary C. Lipton
Homepage https://doi.org/XXXXXXX.XXXXXXX