A learning apparatus and method for creating an emotion expression video according to an embodiment of the present invention comprises: a first generative adversarial network (GAN) which receives an input of a text for creation of an emotion expression video, performs embedding of the input text so as to extract vector information, and creates an image on the basis of the extracted vector information; and a second generative adversarial network which receives an input of an emotion expression image and a frame of a comparative video, and creates a frame of an emotion expression video therefrom.
展开▼