We develop novel algorithms in the area of Pedestrian Attribute Recognition (PAR). Given a person’s image, PAR aims to predict a group of attributes, e.g., gender, age, clothing, color, and attachment to describe the characteristic of this person. While novel tools will be developed, it is not a goal to obtain any results relating to specific persons. Hence no attempts will be made to identify any persons. All image data that could be used to identify people will be deleted as soon as it is not needed for research purposes anymore.
The following publicly available datasets are used for our research:
PETA
http://mmlab.ie.cuhk.edu.hk/projects/PETA.html
PA-100k
https://github.com/xh-liu/HydraPlus-Net/blob/master/README.md
RAP : Richly Annotated Pedestrian Dataset