Face recognition systems based on independent component analysis and support vector machine
Jia Zhang, Yu Shi · 2014
This paper presents an approach for face recognition system based on independent component analysis (ICA) and support vector machine(SVM). The ICA is a feature extraction technique for isolating a multivariate signal into additive subcomponents by considering that the hidden components are non-Gaussian signals. It has been mainly used on the problem of blind signal separation, while support vector machine is a very effective tool to classify the objects/faces into the right category. In this paper, a face recognition system was proposed based on these two techniques. Experiments were carried out on ORL, Yale and YaleB face databases. Simulation results reveal that the proposed system using ICA and SVM can achieve a higher recognition rate with the increasing number of face features. The results also show that the SVM using radial basis functions yields a better performance.