ChatPainter: Improving Text to Image Generation using Dialogue

TL;DR

ChatPainter introduces dialogue-based descriptions to enhance text-to-image generation, significantly improving image quality and understanding of complex scenes in datasets like MS COCO.

Contribution

This work demonstrates that incorporating dialogue descriptions into text prompts improves the fidelity and accuracy of generated images compared to using captions alone.

Findings

Increased inception scores for generated images.

Improved object recognition in synthesized images.

Enhanced scene understanding through dialogue augmentation.

Abstract

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.