Face Hallucination Using Sparse Representation Algorithm

Face Hallucination is a super-resolution technique to obtain high-resolution facial images by taking low-resolution facial images as input. In this paper problem of face hallucination has been approached by using sparse Representation. The image has to be subdivided into different segments so that image pixel information can be retrieved easily from each segment. Each small patch of the image has to be enhanced by using sparse representation. Sparse representation utilizes super resolution model to improve the quality of the image. Super-resolution model couples the small patches dictionaries by using different types of linear shifting experiments. Experimental hallucination results demonstrate that our approach can hallucinate high quality super-resolution faces.