Facial video super resolution using semantic exemplar components | IEEE Conference Publication | IEEE Xplore