Год: 2017
полная версия
We train generative ’up-convolutional’ neural networks which are able to generate images of objects given object style, viewpoint, and color. We train the networks on rendered 3D models of chairs, tables, and cars. Our experiments show that the...
