SeqViews2SeqLabels: Learning 3D Global Features via Aggregating Sequential Views by RNN With Attention | IEEE Journals & Magazine | IEEE Xplore