Generating Images from Captions with Attention

TL;DR

This paper presents an attention-based generative model that creates images from captions, producing higher quality images with novel scene compositions by iteratively drawing patches guided by relevant words.

Contribution

The paper introduces a new attention-driven image generation model that outperforms baselines in quality and scene diversity on captioned datasets.

Findings

Produces higher quality images than baseline models

Generates images with novel scene compositions

Effective in image retrieval tasks

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 Microsoft COCO, we compare our model with several baseline generative models on image generation and retrieval tasks. We demonstrate that our model produces higher quality samples than other approaches and generates images with novel scene compositions corresponding to previously unseen captions in the dataset.