Iterative Prompt Refinement for Safer Text-to-Image Generation

2025-09-17

AI Summary

To address unsafe image generation in text-to-image (T2I) models caused by unsafe prompts, this paper proposes an iterative prompt optimization method grounded in vision-language models (VLMs). Our approach introduces a visual feedback mechanism that jointly leverages textual semantics and generated image content for cross-modal safety assessment, enabling dynamic, intent-aligned prompt refinement. Key contributions include: (1) the first multimodal annotated dataset integrating both textual and visual safety signals; (2) an end-to-end differentiable safety supervision framework supporting VLM-driven prompt rewriting; and (3) comprehensive evaluation across multiple state-of-the-art T2I models, demonstrating substantial improvements in generation safety (+32.7% safety rate) while preserving fidelity and intent alignment (FID improves by 0.8%, CLIP-Score degrades by only 0.02). The implementation is publicly available.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Safety and Robustness
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Security and Privacy: Large-scale security measurements

Abstract

Text-to-Image (T2I) models have made remarkable progress in generating images from text prompts, but their output quality and safety still depend heavily on how prompts are phrased. Existing safety methods typically refine prompts using large language models (LLMs), but they overlook the images produced, which can result in unsafe outputs or unnecessary changes to already safe prompts. To address this, we propose an iterative prompt refinement algorithm that uses Vision Language Models (VLMs) to analyze both the input prompts and the generated images. By leveraging visual feedback, our method refines prompts more effectively, improving safety while maintaining user intent and reliability comparable to existing LLM-based approaches. Additionally, we introduce a new dataset labeled with both textual and visual safety signals using off-the-shelf multi-modal LLM, enabling supervised fine-tuning. Experimental results demonstrate that our approach produces safer outputs without compromising alignment with user intent, offering a practical solution for generating safer T2I content. Our code is available at https://github.com/ku-dmlab/IPR. extbf{ extcolor{red}WARNING: This paper contains examples of harmful or inappropriate images generated by models.

Problem

• Improving text-to-image generation safety through iterative prompt refinement
• Addressing unsafe outputs by analyzing both prompts and generated images
• Maintaining user intent while enhancing visual content safety

Innovation

• Iterative prompt refinement using VLMs
• Leverages visual feedback for safety
• Introduces dataset with multimodal safety labels