A new technique for face image resolution enhancement based on face hallucination

Hallucinate face is an image resolution enhancement algorithm from low-resolution input image to high-resolution based on training example and feature extraction, which high-resolution render image was achieved with the correlative information of training examples. It is important in practice on how to select the sample features and the parameters of recovering algorithm in order to obtain a high quality enhancement. In this paper we propose a new technique to select the feature pyramid of image using pixel down-sampling operator with overlap 3times3-neighbor. Furthermore, the parameters in the recovering algorithm are via solving a combination optimization problem in terms of least squares. The experimental results show that we can get the better vision image with lower RMS errors by using the proposed algorithm.