Subspace analysis for face recognition
Hui Kong · 2006
This thesis presents a research project on face recognition via subspace analysis algorithms. Although face recognition has been actively studied over the past decade, the state-of-the-art recognition systems cannot yield satisfying performance due to small number of training samples available and image variations caused illumination, pose and others. In this thesis, we propose novel subspace analysis approaches that focus on how to recognize human faces with only few training samples.