FAPE-IR: Frequency-Aware Planning and Execution Framework for All-in-One Image Restoration

TL;DR

FAPE-IR introduces a unified, frequency-aware image restoration framework that leverages a multimodal language model for planning and a dynamic expert module for execution, achieving state-of-the-art results across multiple tasks.

Contribution

It presents a novel frequency-aware planning and execution framework using a frozen multimodal language model and a LoRA-MoE module for versatile image restoration.

Findings

Achieves state-of-the-art performance on seven restoration tasks.

Exhibits strong zero-shot generalization under mixed degradations.

Improves restoration quality with adversarial training and frequency regularization.

Abstract

All-in-One Image Restoration (AIO-IR) aims to develop a unified model that can handle multiple degradations under complex conditions. However, existing methods often rely on task-specific designs or latent routing strategies, making it hard to adapt to real-world scenarios with various degradations. We propose FAPE-IR, a Frequency-Aware Planning and Execution framework for image restoration. It uses a frozen Multimodal Large Language Model (MLLM) as a planner to analyze degraded images and generate concise, frequency-aware restoration plans. These plans guide a LoRA-based Mixture-of-Experts (LoRA-MoE) module within a diffusion-based executor, which dynamically selects high- or low-frequency experts, complemented by frequency features of the input image. To further improve restoration quality and reduce artifacts, we introduce adversarial training and a frequency regularization loss. By coupling semantic planning with frequency-based restoration, FAPE-IR offers a unified and interpretable solution for all-in-one image restoration. Extensive experiments show that FAPE-IR achieves state-of-the-art performance across seven restoration tasks and exhibits strong zero-shot generalization under mixed degradations.