Test-Time Preference Optimization for Image Restoration

TL;DR

This paper introduces TTPO, a novel test-time optimization method that improves image restoration quality by aligning outputs with human preferences without retraining models or collecting extensive data.

Contribution

It proposes the first training-free, three-stage test-time preference optimization framework that enhances perceptual quality and adapts to any IR model backbone.

Findings

Effective across various IR tasks and models

Improves perceptual quality aligned with human preferences

Does not require retraining or large preference datasets

Abstract

Image restoration (IR) models are typically trained to recover high-quality images using L1 or LPIPS loss. To handle diverse unknown degradations, zero-shot IR methods have also been introduced. However, existing pre-trained and zero-shot IR approaches often fail to align with human preferences, resulting in restored images that may not be favored. This highlights the critical need to enhance restoration quality and adapt flexibly to various image restoration tasks or backbones without requiring model retraining and ideally without labor-intensive preference data collection. In this paper, we propose the first Test-Time Preference Optimization (TTPO) paradigm for image restoration, which enhances perceptual quality, generates preference data on-the-fly, and is compatible with any IR model backbone. Specifically, we design a training-free, three-stage pipeline: (i) generate candidate preference images online using diffusion inversion and denoising based on the initially restored image; (ii) select preferred and dispreferred images using automated preference-aligned metrics or human feedback; and (iii) use the selected preference images as reward signals to guide the diffusion denoising process, optimizing the restored image to better align with human preferences. Extensive experiments across various image restoration tasks and models demonstrate the effectiveness and flexibility of the proposed pipeline.