1. Ingeniería

TPA 7: Towards Expression invariant
Face Recognition
January 29, 2015
Problem Statement:
expressions.
Input:
The project aims for face recognition under different
Input to the system are the following
• Face image with different expressions for training (neutral expression is
also included).
Expected Output:
The developed code should be able to do the following
• Given a probe face image, the person must be identified.
• ROC, CMS curves, Confusion matrix plots.
Hint for excellence:
Special credit will be given if a modified or adaptive
FR system is designed to deal with different expressions and if the system can
identify the expression, also if pose and occlusion can be handled.
References
• Zhen Lei; Shengcai Liao; Pietikainen, M.; Li, S.Z.; , ”Face Recognition
by Exploring Information Jointly in Space, Scale and Orientation,” IEEE
Transactions on Image Processing, vol.20, no.1, pp.247-256, Jan. 2011
• Salazar, Augusto, et al. ”Fully automatic expression-invariant face correspondence.” Machine Vision and Applications 25.4 (2014): 859-879.
• Santhanaganesh, A. S., and P. S. Rajakumar. ”Facial Expression Recognition in Various Illuminous Environment.” Digital Image Processing 6.3
(2014): 127-130.
• Salazar, Augusto, et al. ”Fully automatic expression-invariant face correspondence.” Machine Vision and Applications 25.4 (2014): 859-879.
1
• Bronstein, Alexander M., Michael M. Bronstein, and Ron Kimmel. ”Robust expression-invariant face recognition from partially missing data.”
Computer VisionECCV 2006. Springer Berlin Heidelberg, 2006. 396-408.
• Amberg, Brian, Reinhard Knothe, and Thomas Vetter. ”Expression invariant 3D face recognition with a morphable model.” Automatic Face &
Gesture Recognition, 2008. 8th IEEE International Conference on. IEEE,
2008.
2