Expression-invariant Non-rigid 3D Face Recognition: A Robust Approach to Expression-aware Morphing
Faisal R. Al-Osaimi, Mohammed Bennamoun, Ajmal Saeed Mian · UWA Profiles and Research Repository (University of Western Australia) · 2008
An automatic approach to the recognition of 3D face under any facial expression is presented. It has the capacity to differentiate between expression deformations, those which are caused by expressions, and interpersonal disparities that are essential for recognition. When two facial scans are matched the expression deformations are morphed out before similarity measure calculation while the interpersonal disparities are retained. A combination of PCA and ICP is used in our system. The PCA subspace is built in such a way it models the facial expressions while it does not model the human face. The approach was applied on the FRGC v2.0 dataset and superior recognition performance was achieved. The verification rates at 0.001FAR were 98.35% and 97.73% for scans under neutral and non-neutral expressions, respectively.