From pixels to gestures: learning visual representations for human analysis in color and depth data sequences | Publicación