A new framework to combine vertical and horizontal information for face recognition

We present an efficient feature extraction framework named two-dimensional combined complex discriminant analysis (2DCCDA) for face recognition. In this framework, 2DLDA is performed vertically and, at the same time, horizontally. That is, both vertical and horizontal discriminant information can be extracted separately. Then both the horizontal and the vertical feature matrices are combined into a complex feature matrix. A complex version of 2DLDA is introduced to extract the discriminant complex features of this complex feature matrix for feature selection. Experiments on the AT&T and AR databases show that 2DCCDA achieves satisfactory performance not only under the conditions of varied facial expression and lighting configuration but also under the conditions where the pose and sample size are varied.