Face detection by using independent component decomposition
Ioan Buciu, Constantine L. Kotropoulos, Ioannis Pitas · 2002
In this paper we explore the independent component decomposition for face detection. The minimization of the Kullback Leibler divergence and the maximization of the entropy are two methods employed to decompose an original image into its independent components. We built nearest neighbor classifiers based on their resulting independent components and compare their ability to detect faces to that of support vector machines.