A Fine-Grained Spatial-Temporal Attention Model for Video Captioning | IEEE Journals & Magazine | IEEE Xplore