ICA Based on KPCA and Hybrid Flexible Neural Tree to Face Recognition

Abstract

In this paper, a new approach using independent component analysis (ica) and hybrid Flexible Neural Tree (FNT) is put forward for face recognition. To improve the quality of the face images, a series of image pre-processing techniques, which include histogram equalization, edge detection and geometrical transformation are used. The ICA based on Kernel principal component analysis (KPCA) and FastICA is employed to extract features, and the Hybrid FNT is used to identify the faces. To accelerate the convergence of the FNT and improve the quality of the solutions, the extended compact genetic programming (ECGP) and particle swarm optimization (PSO) are applied to optimize the FNT structure and parameters. The experimental results show that the proposed framework is efficient for face recognition.