Spatio-Temporal Encoder-Decoder Fully Convolutional Network for Video-Based Dimensional Emotion Recognition | IEEE Journals & Magazine | IEEE Xplore