ML Anthology

A Picture Is Worth a Thousand Prompts? Efficacy of Iterative Human-Driven Prompt Refinement in Image Regeneration Tasks

Khoi Trinh, Scott Seidenberger, Raveen Wijewickrama, Murtuza Jadliwala, Anindya Maiti
IJCAI 2025 pp. 10189-10197
doi:10.24963/IJCAI.2025/1132 /ijcai/2025/trinh2025ijcai-picture/

Abstract

With AI-generated content becoming widespread across digital platforms, it is important to understand how such content is inspired and produced. This study explores the underexamined task of image regeneration, where a human operator iteratively refines prompts to recreate a specific target image. Unlike typical image generation, regeneration begins with a visual reference. A key challenge is whether existing image similarity metrics (ISMs) align with human judgments and can serve as useful feedback in this process. We conduct a structured user study to evaluate how iterative prompt refinement affects similarity to target images and whether ISMs reflect the improvements perceived by human observers. Our results show that prompt adjustments significantly improve alignment, both subjectively and quantitatively, highlighting the potential of iterative workflows in enhancing generative image quality.

PDF IJCAI Semantic Scholar

Cite

Text

Trinh et al. "A Picture Is Worth a Thousand Prompts? Efficacy of Iterative Human-Driven Prompt Refinement in Image Regeneration Tasks." International Joint Conference on Artificial Intelligence, 2025. doi:10.24963/IJCAI.2025/1132

Markdown

[Trinh et al. "A Picture Is Worth a Thousand Prompts? Efficacy of Iterative Human-Driven Prompt Refinement in Image Regeneration Tasks." International Joint Conference on Artificial Intelligence, 2025.](https://mlanthology.org/ijcai/2025/trinh2025ijcai-picture/) doi:10.24963/IJCAI.2025/1132

BibTeX

@inproceedings{trinh2025ijcai-picture,
  title     = {{A Picture Is Worth a Thousand Prompts? Efficacy of Iterative Human-Driven Prompt Refinement in Image Regeneration Tasks}},
  author    = {Trinh, Khoi and Seidenberger, Scott and Wijewickrama, Raveen and Jadliwala, Murtuza and Maiti, Anindya},
  booktitle = {International Joint Conference on Artificial Intelligence},
  year      = {2025},
  pages     = {10189-10197},
  doi       = {10.24963/IJCAI.2025/1132},
  url       = {https://mlanthology.org/ijcai/2025/trinh2025ijcai-picture/}
}