Improving Text-to-Image Consistency via Automatic Prompt Optimization

TL;DR

This paper introduces OPT2I, a prompt optimization framework using large language models to enhance text-to-image consistency without fine-tuning models, validated on multiple datasets with significant improvements.

Contribution

The paper proposes a novel LLM-based prompt optimization method that improves T2I consistency, addressing limitations of existing prompt refinement approaches.

Findings

Boosts consistency score by up to 24.9% on datasets

Preserves image quality as measured by FID

Increases recall between generated and real data

Abstract

Impressive advances in text-to-image (T2I) generative models have yielded a plethora of high performing models which are able to generate aesthetically appealing, photorealistic images. Despite the progress, these models still struggle to produce images that are consistent with the input prompt, oftentimes failing to capture object quantities, relations and attributes properly. Existing solutions to improve prompt-image consistency suffer from the following challenges: (1) they oftentimes require model fine-tuning, (2) they only focus on nearby prompt samples, and (3) they are affected by unfavorable trade-offs among image quality, representation diversity, and prompt-image consistency. In this paper, we address these challenges and introduce a T2I optimization-by-prompting framework, OPT2I, which leverages a large language model (LLM) to improve prompt-image consistency in T2I models. Our framework starts from a user prompt and iteratively generates revised prompts with the goal of maximizing a consistency score. Our extensive validation on two datasets, MSCOCO and PartiPrompts, shows that OPT2I can boost the initial consistency score by up to 24.9% in terms of DSG score while preserving the FID and increasing the recall between generated and real data. Our work paves the way toward building more reliable and robust T2I systems by harnessing the power of LLMs.