NIR-to-VIS Face Recognition via Embedding Relations and Coordinates of the Pairwise Features | Read Paper on Bytez