Description

13,000 cropped facial regions (using; Viola-Jones that have been labeled with a name identifier. A subset of the people present have two images in the datasetits quite common for people to train facial matching systems here.

Related datasets