PCR

Research on Attributes

Attributes dependencies
Dependencies between facial attributes given the appearance. Recent works report significant drops in performance for state-of-the-art gender classifiers when evaluated "in the wild", i.e. with uncontrolled demography and environmental conditions. We hypothesize that this is caused by the existence of dependencies among facial demographic attributes that have not been considered when building the classifier. In the paper we study the dependencies among gender, age and pose facial attributes. By considering the relation between gender and pose attributes we also avoid the use of computation-ally expensive and fragile face alignment procedures. In the experiments we confirm the existence of dependencies among gender, age and pose facial attributes and prove that we can improve the performance and robustness of gender classifiers by exploiting these dependencies.
Related Publications: CIARP'2011, IBPRIA'2013, PRL'2014
Age estimation
Facial age estimation . We developed a "light weight" face age estimation algorithm using PCA+LDA and KNN regression.
Related Publications: IBPRIA'2011,
Gender recognition
PCA+LDA face gender recognition. We developed gender classification algorithm using an simple holistic approach. We got 93% accuracy on FERET database (5-fold cross validation) and an speed comparable to the fastest gender recognition algorithms. This algorithm is based on PCA+LDA with proper cross-validation of PCA dimensions.
Related Publications: PAMI'2011,
Face expressions recognition
Face expressions recognition. We used our 2D appearance tracking algorithm [BMVC'2006] to track the face. We developed a user independent manifold of face expressions using PCA+LDA and then we classified the expressions using KNN.
Related Publications: PAA'2008, FG2008.
Appearance-based 3D reanimation
Appearance-based reanimation from expressions estimation. We developed a way of doing 3D face reanimation based on appearance-based technicques.
Related Publications: IbPRIA'2005. ICPR'2006.
Some videos: [video 1], [video 2], [video 3], [video 4], [video 5]