Fast multi-view face alignment via multi-task auto-encoders | IEEE Conference Publication | IEEE Xplore