Pose-Robust and Discriminative Feature Representation by Multi-task Deep Learning for Multi-view Face Recognition | IEEE Conference Publication | IEEE Xplore