Robust Facial Attribute-Specific Subspace-Based Principal Component Analysis for Face Recognition

Facial attribute-specific subspace-based PCA (FASS-based PCA) considers the information of class labels, and the discriminant power can be improved. However, it doesn´t consider the outliers which are .common in realistic training sets. To address this problem, we propose robust facial attribute-specific subspace-based PCA (robust FASS-based PCA) algorithm in this paper, which gives a new weighted method for the evaluation of the total squared error of each class. The detailed algorithm for robust FASS-based PCA is also given. The results of experiments conducted on Yale databases show the effectiveness of the new feature extraction algorithm.