Abstract

Abstract Motivated by the recent progress in generative models, we introduce a model that generates images from natural language descriptions. The proposed model iteratively draws patches on a canvas, while attending to the relevant words in the description. After training on MS COCO, we compare our models with several baseline generative models on image generation and retrieval tasks. We demonstrate our model produces higher quality samples than other approaches and generates images with novel scene compositions corresponding to previously unseen captions in the dataset. For more details, visit http://arxiv.org/abs/ 1511.02793.