A Preliminary Study on GPT-Image Generation Model for Image Restoration

TL;DR

This paper systematically evaluates GPT-Image models for image restoration, revealing their perceptual quality, limitations in structural fidelity, and potential as priors to enhance existing restoration methods across various scenarios.

Contribution

It provides the first benchmark of GPT-Image in restoration tasks, analyzes their strengths and weaknesses, and demonstrates their utility as priors to improve restoration performance.

Findings

GPT-Image produces perceptually pleasant images but lacks pixel-level fidelity.

Incorporating GPT-generated priors improves restoration quality in case studies.

The study offers a baseline framework for integrating GPT models into restoration pipelines.

Abstract

Recent advances in OpenAI's GPT-series multimodal generation models have shown remarkable capabilities in producing visually compelling images. In this work, we investigate its potential impact on the image restoration community. We provide, to the best of our knowledge, the first systematic benchmark across diverse restoration scenarios. Our evaluation shows that, while the restoration results generated by GPT-Image models are often perceptually pleasant, they tend to lack pixel-level structural fidelity compared with ground-truth references. Typical deviations include changes in image geometry, object positions or counts, and even modifications in perspective. Beyond empirical observations, we further demonstrate that outputs from GPT-Image models can act as strong visual priors, offering notable performance improvements for existing restoration networks. Using dehazing, deraining, and low-light enhancement as representative case studies, we show that integrating GPT-generated priors significantly boosts restoration quality. This study not only provides practical insights and a baseline framework for incorporating GPT-based generative priors into restoration pipelines, but also highlights new opportunities for bridging image generation models and restoration tasks. To support future research, we will release GPT-restored results.