Selective coding of human faces using wavelets

In this paper, we propose an automated selective coding algorithm for human faces using neural networks and wavelets. In the proposed coding algorithm, we try to preserve information of human faces as much as possible without compromising overall compression efficiency. In particular, we want to eliminate the artifacts near the boundary between the face and background. We first extract the facial area using the location information of the eye and the skin color information. When we allocate bits at each level in wavelet transform, we allocate more bits to the area corresponding to the facial area. Experiments show that we can obtain a crisper facial area at the expense of the background.