GR-GAN: Gradual Refinement Text-to-image Generation

TL;DR

GR-GAN is a novel text-to-image generation model that progressively refines images from coarse to fine stages, ensuring high quality and text-image consistency, and introduces a new evaluation metric.

Contribution

The paper introduces GR-GAN, a progressive refinement framework with stage-wise text constraints and a new metric, CMD, for improved image quality and consistency in text-to-image synthesis.

Findings

Outperforms previous models on FID and CMD metrics

Achieves state-of-the-art results in image quality and text-image alignment

Demonstrates efficiency of multi-stage generation process

Abstract

A good Text-to-Image model should not only generate high quality images, but also ensure the consistency between the text and the generated image. Previous models failed to simultaneously fix both sides well. This paper proposes a Gradual Refinement Generative Adversarial Network (GR-GAN) to alleviates the problem efficiently. A GRG module is designed to generate images from low resolution to high resolution with the corresponding text constraints from coarse granularity (sentence) to fine granularity (word) stage by stage, a ITM module is designed to provide image-text matching losses at both sentence-image level and word-region level for corresponding stages. We also introduce a new metric Cross-Model Distance (CMD) for simultaneously evaluating image quality and image-text consistency. Experimental results show GR-GAN significant outperform previous models, and achieve new state-of-the-art on both FID and CMD. A detailed analysis demonstrates the efficiency of different generation stages in GR-GAN.