ChatPainter: Improving Text to Image Generation using Dialogue

Shikhar Sharma, Dendi Suhubdy, Vincent Michalski, Samira Ebrahimi Kahou, Yoshua Bengio


Synthesizing realistic images from text descriptions on a dataset like Microsoft Common Objects in Context (MS COCO), where each image can contain several objects, is a challenging task. Prior work has used text captions to generate images. However, captions might not be informative enough to capture the entire image and insufficient for the model to be able to understand which objects in the images correspond to which words in the captions. We show that adding a dialogue that further describes the scene leads to significant improvement in the inception score and in the quality of generated images on the MS COCO dataset.


  title     = {{C}hat{P}ainter: Improving Text to Image Generation using Dialogue},
  author    = {Shikhar Sharma and
               Dendi Suhubdy and
               Vincent Michalski and
               Samira Ebrahimi Kahou and
               Yoshua Bengio},
  booktitle = {International Conference on Learning Representations (ICLR) Workshop},
  month     = {April},
  year      = {2018},
  url       = {}