PIV-FlowDiffuser:Transfer-learning-based denoising diffusion models for PIV

2025-04-21

AI Summary

To address residual vector-field noise in particle image velocimetry (PIV) caused by domain shift between synthetic and real particle images, this paper proposes FlowDiffuser—the first denoising diffusion probabilistic model (DDPM) tailored for PIV. Methodologically, it pioneers the adaptation of optical flow–pretrained diffusion models to PIV; introduces a two-stage transfer learning paradigm (Sintel/KITTI → synthetic PIV → real PIV); and incorporates 2× super-resolution upsampling to resolve small-scale turbulent structures. Technically, it integrates DDPM with a RAFT256-PIV baseline and domain-adaptive fine-tuning. Evaluated on Cai’s dataset, FlowDiffuser achieves a 59.4% reduction in average endpoint error (AEE), effectively suppresses domain-specific noise, and demonstrates substantially improved generalization across diverse PIV scenarios.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Search and Optimization: Learning to Search

Application Category

• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Deep learning algorithms have significantly reduced the computational time and improved the spatial resolution of particle image velocimetry~(PIV). However, the models trained on synthetic datasets might have a degraded performance on practical particle images due to domain gaps. As a result, special residual patterns are often observed for the vector fields of deep learning-based estimators. To reduce the special noise step-by-step, we employ a denoising diffusion model~(FlowDiffuser) for PIV analysis. And the data-hungry iterative denoising diffusion model is trained via a transfer learning strategy, resulting in our PIV-FlowDiffuser method. Specifically, (1) pre-training a FlowDiffuser model with multiple optical flow datasets of the computer vision community, such as Sintel, KITTI, etc; (2) fine-tuning the pre-trained model on synthetic PIV datasets. Note that the PIV images are upsampled by a factor of two to resolve the small-scale turbulent flow structures. The visualized results indicate that our PIV-FlowDiffuser effectively suppresses the noise patterns. Therefore, the denoising diffusion model reduces the average end-point error~($AEE$) by 59.4% over RAFT256-PIV baseline on the classic Cai's dataset. Besides, PIV-FlowDiffuser exhibits enhanced generalization performance on unseen particle images due to transfer learning. Overall, this study highlights the transfer-learning-based denoising diffusion models for PIV. And a detailed implementation is recommended for interested readers in the repository https://github.com/Zhu-Qianyu/PIV-FlowDiffuser.

Problem

• Reducing domain gap noise in PIV using diffusion models
• Enhancing PIV accuracy via transfer-learning-based denoising
• Improving generalization on unseen particle images with FlowDiffuser

Innovation

• Transfer-learning-based denoising diffusion model
• Pre-training with multiple optical flow datasets
• Fine-tuning on synthetic PIV datasets