PrefPaint: Aligning Image Inpainting Diffusion Model with Human Preference

TL;DR

PrefPaint introduces a reinforcement learning framework that aligns image inpainting diffusion models with human aesthetic preferences, significantly enhancing visual quality and appeal.

Contribution

It is the first to incorporate human preference into diffusion model fine-tuning for image inpainting using a reward model and theoretical error bounds.

Findings

Improved inpainting quality aligned with human preferences

Effective reinforcement learning-based fine-tuning of diffusion models

Enhanced performance in downstream tasks like image extension and 3D reconstruction

Abstract

In this paper, we make the first attempt to align diffusion models for image inpainting with human aesthetic standards via a reinforcement learning framework, significantly improving the quality and visual appeal of inpainted images. Specifically, instead of directly measuring the divergence with paired images, we train a reward model with the dataset we construct, consisting of nearly 51,000 images annotated with human preferences. Then, we adopt a reinforcement learning process to fine-tune the distribution of a pre-trained diffusion model for image inpainting in the direction of higher reward. Moreover, we theoretically deduce the upper bound on the error of the reward model, which illustrates the potential confidence of reward estimation throughout the reinforcement alignment process, thereby facilitating accurate regularization. Extensive experiments on inpainting comparison and downstream tasks, such as image extension and 3D reconstruction, demonstrate the effectiveness of our approach, showing significant improvements in the alignment of inpainted images with human preference compared with state-of-the-art methods. This research not only advances the field of image inpainting but also provides a framework for incorporating human preference into the iterative refinement of generative models based on modeling reward accuracy, with broad implications for the design of visually driven AI applications. Our code and dataset are publicly available at https://prefpaint.github.io.