FideDiff: Efficient Diffusion Model for High-Fidelity Image Motion Deblurring

2025-10-01

AI Summary

Existing diffusion models for motion deblurring suffer from slow inference and low reconstruction fidelity. This paper proposes FideDiff—the first single-step, high-fidelity diffusion-based deblurring framework. To address these limitations, we (1) synthesize blur-trajectory-aware training data and formulate deblurring as a conditional diffusion process; (2) introduce a consistency training strategy coupled with adaptive timestep prediction to enhance reconstruction stability; and (3) integrate a Kernel-ControlNet module to enable blur-kernel-aware conditional generation. Quantitatively, FideDiff achieves state-of-the-art performance among diffusion-based methods on full-reference metrics (e.g., PSNR, SSIM), matching the fidelity of top non-diffusion approaches while accelerating inference by two orders of magnitude. Our work establishes a new paradigm for high-quality image restoration that simultaneously delivers both computational efficiency and reconstruction accuracy.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Kernel Methods
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Recent advancements in image motion deblurring, driven by CNNs and transformers, have made significant progress. Large-scale pre-trained diffusion models, which are rich in true-world modeling, have shown great promise for high-quality image restoration tasks such as deblurring, demonstrating stronger generative capabilities than CNN and transformer-based methods. However, challenges such as unbearable inference time and compromised fidelity still limit the full potential of the diffusion models. To address this, we introduce FideDiff, a novel single-step diffusion model designed for high-fidelity deblurring. We reformulate motion deblurring as a diffusion-like process where each timestep represents a progressively blurred image, and we train a consistency model that aligns all timesteps to the same clean image. By reconstructing training data with matched blur trajectories, the model learns temporal consistency, enabling accurate one-step deblurring. We further enhance model performance by integrating Kernel ControlNet for blur kernel estimation and introducing adaptive timestep prediction. Our model achieves superior performance on full-reference metrics, surpassing previous diffusion-based methods and matching the performance of other state-of-the-art models. FideDiff offers a new direction for applying pre-trained diffusion models to high-fidelity image restoration tasks, establishing a robust baseline for further advancing diffusion models in real-world industrial applications. Our dataset and code will be available at https://github.com/xyLiu339/FideDiff.

Problem

• Addressing slow inference time in diffusion models
• Improving image fidelity in motion deblurring tasks
• Enabling single-step processing for efficient deblurring

Innovation

• Single-step diffusion model for efficient deblurring
• Consistency model aligns blurred timesteps to clean image
• Kernel ControlNet integration for blur trajectory estimation