EvBS: Event-guided Blur Synthesis for Domain-adaptive Motion Deblurring

2026-08-08

AI Summary

This work addresses the performance degradation of existing pre-trained deblurring models in real-world scenarios due to domain shift. The authors propose EvBS, a novel framework that leverages the high temporal resolution of event cameras to disentangle motion dynamics from visual content—a coupling inherently present in conventional intensity images. To enhance generalization, the method employs both intrinsic and extrinsic blur synthesis strategies to generate diverse training data and enables target-domain-adaptive fine-tuning. Notably, EvBS supports cross-sample motion transfer, significantly improving robustness and deblurring performance on unseen test sets. Extensive experiments demonstrate that the proposed approach outperforms current state-of-the-art methods across multiple benchmarks.

Abstract

Motion deblurring has achieved remarkable progress with deep learning, yet pre-trained deblurring models often suffer from performance degradation in real-world scenarios due to the domain shift between training and testing distributions. To remedy this, we propose EvBS, an event-guided blur synthesis framework that generates diverse training pairs for calibrating pre-trained models to the target domain. While existing methods are constrained by the inherent entanglement between motion and visual content, our method leverages the high temporal resolution of event cameras to effectively decouple them. This enables us to utilize not only the intrinsic motion that is inherent to the given content but also extrinsic motion transferred from different sources within the target domain, thereby facilitating effective adaptation via fine-tuning. Specifically, EvBS comprises two complementary strategies: Intrinsic-Blur Synthesis, which blurs sharp contents with their own motion patterns, and Extrinsic-Blur Synthesis, which transfers motion from blurry patches to distinct sharp content. This approach generates a diverse set of training pairs that break the inherent constraints of naturally coupled motion and content, resulting in enhanced domain-adaptive deblurring performance. Extensive experiments on multiple benchmarks demonstrate that EvBS effectively enhances the robustness of existing deblurring models on unseen testing datasets.

Problem

• motion deblurring
• domain shift
• domain adaptation
• blur synthesis
• event cameras

Innovation

• event camera
• blur synthesis
• domain adaptation
• motion deblurring
• motion-content decoupling