From Few To Many: Illumination Cone Models For Face Recognition Under Variable Lighting And Pose | Manumap