MCFormer: A Multi-Cost-Volume Network and Comprehensive Benchmark for Particle Image Velocimetry

2025-07-07

AI Summary

The particle image velocimetry (PIV) community lacks a standardized optical flow evaluation benchmark. Method: This work introduces the first large-scale synthetic PIV dataset generated via high-fidelity computational fluid dynamics (CFD) simulation and establishes a comprehensive, PIV-specific benchmarking platform. We further propose MCFormer, a multi-frame optical flow network tailored for PIV: it employs a multi-cost-volume architecture to explicitly model particle sparsity and continuous motion, and integrates Transformer-based modules to capture long-range temporal dependencies across frames. Contribution/Results: On the proposed benchmark, MCFormer achieves the lowest normalized end-point error (NEPE), significantly outperforming state-of-the-art optical flow methods. The open-sourced dataset, implementation code, and benchmarking protocol provide a standardized foundation for fair, reproducible evaluation and accelerated deep learning research in PIV.

Technology Category

Computer Vision: Motion & TrackingSearch and Optimization: Sampling/Simulation-based SearchPlanning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

Security and Privacy: Large-scale security measurementsSystems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterizationUser Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Particle Image Velocimetry (PIV) is fundamental to fluid dynamics, yet deep learning applications face significant hurdles. A critical gap exists: the lack of comprehensive evaluation of how diverse optical flow models perform specifically on PIV data, largely due to limitations in available datasets and the absence of a standardized benchmark. This prevents fair comparison and hinders progress. To address this, our primary contribution is a novel, large-scale synthetic PIV benchmark dataset generated from diverse CFD simulations (JHTDB and Blasius). It features unprecedented variety in particle densities, flow velocities, and continuous motion, enabling, for the first time, a standardized and rigorous evaluation of various optical flow and PIV algorithms. Complementing this, we propose Multi Cost Volume PIV (MCFormer), a new deep network architecture leveraging multi-frame temporal information and multiple cost volumes, specifically designed for PIV's sparse nature. Our comprehensive benchmark evaluation, the first of its kind, reveals significant performance variations among adapted optical flow models and demonstrates that MCFormer significantly outperforms existing methods, achieving the lowest overall normalized endpoint error (NEPE). This work provides both a foundational benchmark resource essential for future PIV research and a state-of-the-art method tailored for PIV challenges. We make our benchmark dataset and code publicly available to foster future research in this area.

Problem

• Lack of standardized benchmark for PIV optical flow evaluation
• Need for diverse synthetic PIV dataset from CFD simulations
• Improving deep learning performance on sparse PIV data

Innovation

• Large-scale synthetic PIV benchmark dataset
• Multi Cost Volume PIV (MCFormer) network
• Leverages multi-frame temporal information