From Sharp to Blur: Unsupervised Domain Adaptation for 2D Human Pose Estimation Under Extreme Motion Blur Using Event Cameras

2025-07-30

AI Summary

Extreme motion blur severely degrades 2D human pose estimation performance. To address this, we propose an unsupervised domain adaptation method leveraging event cameras. Exploiting their high temporal resolution, we design a motion-aware data augmentation strategy to synthesize photorealistic motion-blurred images. We further introduce a student–teacher framework with iterative pseudo-label refinement guided by mutual uncertainty masking, effectively mitigating inter-domain distribution shift. Our approach requires no annotations in the target (blurred) domain and jointly integrates event-driven augmentation, cross-domain knowledge distillation, and uncertainty modeling to enable robust transfer from sharp source-domain images to motion-blurred target-domain images. Extensive experiments on multiple motion-blur pose benchmarks demonstrate significant improvements over existing unsupervised domain adaptation methods, validating the unique value and practical potential of event cameras in real-world dynamic blur scenarios.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Intelligent Robots: State Estimation

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Human pose estimation is critical for applications such as rehabilitation, sports analytics, and AR/VR systems. However, rapid motion and low-light conditions often introduce motion blur, significantly degrading pose estimation due to the domain gap between sharp and blurred images. Most datasets assume stable conditions, making models trained on sharp images struggle in blurred environments. To address this, we introduce a novel domain adaptation approach that leverages event cameras, which capture high temporal resolution motion data and are inherently robust to motion blur. Using event-based augmentation, we generate motion-aware blurred images, effectively bridging the domain gap between sharp and blurred domains without requiring paired annotations. Additionally, we develop a student-teacher framework that iteratively refines pseudo-labels, leveraging mutual uncertainty masking to eliminate incorrect labels and enable more effective learning. Experimental results demonstrate that our approach outperforms conventional domain-adaptive human pose estimation methods, achieving robust pose estimation under motion blur without requiring annotations in the target domain. Our findings highlight the potential of event cameras as a scalable and effective solution for domain adaptation in real-world motion blur environments. Our project codes are available at https://github.com/kmax2001/EvSharp2Blur.

Problem

• Addressing motion blur in human pose estimation using event cameras
• Bridging domain gap between sharp and blurred images without paired annotations
• Improving pose estimation accuracy in low-light and rapid motion conditions

Innovation

• Event cameras for motion blur robustness
• Event-based augmentation bridges domain gap
• Student-teacher framework refines pseudo-labels