A. Gallagher and T. Chen
In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 256-263, 2009. The dataset consists of Flick face images labeled with 7 age categories.