Maximum A Posteriori Estimation of Distances Between Deep Features in Still-to-Video Face Recognition | Read Paper on Bytez