Video Captioning Via Two-stage Attention Model and Generative Adversarial Network | IEEE Conference Publication | IEEE Xplore