Face verification in complex background based on imperialist competitive algorithm

Verifying human faces in a gray scale image with complex background is a challenging task since every position and size can be a potential face in these images. Searching for face of interest using regular search methods is an extremely time consuming task. In order to deal with this large search space, this paper proposes the use of Imperialist Competitive Algorithm (ICA). ICA is an evolutionary algorithm for global optimization. In the proposed method ICA selects sub-images with different sizes from an input image heuristically. The fitness of each sub-image is calculated using the eigenfaces method. The fitness is composed of two terms: The first term is the distance between the projection of sub-image to the eigen space and the closest projection of images in the training set and the second term is the cross-correlation between the original sub-image and the reconstructed sub-image using eigenfaces. Experimental results show a high correct verification rate for images from different scenes.