Labeled Fiducial Points in Faces for Aging and Expression Study

Fiducial points are manually labeled in face images from two databases: Lifespan and FACES.

Our manually labeled fiducial points can be used for research purpose only, such as facial expression recognition with aging variations [1]. The face images in these two databases (see references in [1] for introduction of the two databases originally captured by psychology researchers) can also be used for age estimation across facial expression changes [2, 3]. 

Please cite the related papers from us if our labeled fiducial points are used in your research publications.

The files can be downloaded from here: Lifespan_Fiducials.rar, FACES_Fiducials.rar 

The fiducial points are stored in .mat files that can be viewed from Matlab.

The file is protected by a password. Please email me (guodong dot guo AT mail dot wvu dot edu) to request the passwords to uncompress the two files.

References:

[1]  G-D. Guo, R. Guo, and X. Li, Facial expression recognition influenced by human aging, IEEE Trans. Affective Computing, vol. 4, no. 3, pp. 291-298, July-Sept. 2013.

[2]  G-D. Guo and X. Wang, A Study on Human Age Estimation under Facial Expression Changes, IEEE Conference on Computer Vision and Pattern Recognition (CVPR'12), June, 2012.

[3]  C. Zhang and G-D. Guo, Age estimation with expression changes using multiple aging subspaces, In the IEEE Sixth International Conference on Biometrics: Theory, Applications and Systems (BTAS 2013), Washington DC, September, 2013.