Implementation of incremental linear discriminant analysis using singular value decomposition for face recognition

Dimensionality reduction methods have been successfully employed for face recognition. Among the various dimensionality reduction algorithms, linear discriminant analysis (LDA) is one of the popular supervised dimensionality reduction methods, and many LDA-based face recognition algorithms/systems have been reported in the last decade. However, the LDA based face recognition systems suffer from the scalability problem. To overcome this limitation, an incremental approach is usually adopted. Incremental LDA (ILDA) methods have been studied for many years in the statistical and machine-learning community. The major limitation of existing ILDA method is to handle the inverse of the within-class scatter matrix. In view of this limitation, this paper proposes a new ILDA method based on the idea of singular value decomposition (SVD) updating algorithm, namely an SVD updating-based ILDA (ILDA-SVD) algorithm. In the proposed ILDA-SVD algorithm, it is proved that the approximation error is mathematically bounded. The proposed method has been evaluated using available public databases, namely Yale, Yale B and ORL are applied to existing face-recognition algorithms. The proposed methods are successfully applied to face-recognition, and the simulation results on Yale database show that the proposed algorithms significantly outperform other well-known systems in terms of recognition rate.