Short-term anchor linking and long-term self-guided attention for video object detection | Publicación