Region-based reconstruction for face hallucination

This paper proposes a new method for synthesizing high-resolution faces from single-frame low-resolution facial images, using a region-based reconstruction method of an extended morphable face model. In the method, we suppose that any noble facial image can be reconstructed by a linear combination of prototypes obtained from training facial images. Then, in order to maintain the local characteristics of local facial regions, we apply a region-based reconstruction method. The encouraging results show that the proposed methods can be used to improve the performance of face recognition systems, particularly to enhance the resolution of facial images captured from visual surveillance systems.