Abstract
We train a generative convolutional neural network which is able to generate images of objects given object type,viewpoint,and color. We train the network in a su-pervised manner on a dataset of rendered 3D chair mod-els.Our experiments show that the nerwork does not merely learn all images by heart,but rather finds a meaningfiul representation of a 3D chair model allowing it to assess the similariry of different chairs,interpolate between given viewpoints to generate the missing ones,or invent new chair styles by interpolating between chairs from the training set.We show that the network can be used to find correspon-dences between diferent chairs from the dataset,outper-forming existing approaches on this task.