Global Spatial-Temporal Information Encoder-Decoder Based Action Segmentation in Untrimmed Video | TUP Journals & Magazine | IEEE Xplore