SurfPhase: 3D Interfacial Dynamics in Two-Phase Flows from Sparse Videos

📅 2026-02-11
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
Accurately capturing the three-dimensional dynamics of deformable liquid–gas interfaces in two-phase flows remains challenging, particularly under sparse observational views and sharp interfacial boundaries. This work proposes a novel approach that, for the first time, integrates dynamic Gaussian surfels with signed distance functions to enforce geometric consistency, while leveraging a video diffusion model to synthesize high-quality novel views from extremely sparse inputs—such as only two viewpoints—to enhance 3D interface reconstruction. Evaluated on a high-speed pool boiling dataset, the method achieves high-fidelity view synthesis and accurate interfacial velocity estimation, significantly improving the precision and robustness of modeling two-phase flow interface dynamics under sparse observational conditions.

Technology Category

Application Category

📝 Abstract
Interfacial dynamics in two-phase flows govern momentum, heat, and mass transfer, yet remain difficult to measure experimentally. Classical techniques face intrinsic limitations near moving interfaces, while existing neural rendering methods target single-phase flows with diffuse boundaries and cannot handle sharp, deformable liquid-vapor interfaces. We propose SurfPhase, a novel model for reconstructing 3D interfacial dynamics from sparse camera views. Our approach integrates dynamic Gaussian surfels with a signed distance function formulation for geometric consistency, and leverages a video diffusion model to synthesize novel-view videos to refine reconstruction from sparse observations. We evaluate on a new dataset of high-speed pool boiling videos, demonstrating high-quality view synthesis and velocity estimation from only two camera views. Project website: https://yuegao.me/SurfPhase.
Problem

Research questions and friction points this paper is trying to address.

two-phase flows
interfacial dynamics
liquid-vapor interface
3D reconstruction
sparse video
Innovation

Methods, ideas, or system contributions that make the work stand out.

two-phase flow
3D reconstruction
Gaussian surfels
signed distance function
video diffusion model
🔎 Similar Papers
No similar papers found.