Cross-View Action Recognition Using Contextual Maximum Margin Clustering | IEEE Journals & Magazine | IEEE Xplore