In text-to-image generation, using negative prompts, which describe undesirable image characteristics, can significantly boost image quality. However, producing good negative prompts is manual and tedious. To address this, we propose NegOpt, a novel method for optimizing negative prompt generation toward enhanced image generation, using supervised fine-tuning and reinforcement learning. Our combined approach results in a substantial increase of 25% in Inception Score compared to other approaches and surpasses ground-truth negative prompts from the test set. Furthermore, with NegOpt we can preferentially optimize the metrics most important to us. Finally, we construct Negative Prompts DB (https://huggingface.co/datasets/mikeogezi/negopt_full), a publicly available dataset of negative prompts.
@article{arxiv.2403.07605,
title = {Optimizing Negative Prompts for Enhanced Aesthetics and Fidelity in Text-To-Image Generation},
author = {Michael Ogezi and Ning Shi},
journal= {arXiv preprint arXiv:2403.07605},
year = {2024}
}