Snap-and-tune: combining deep learning and test-time optimization for high-fidelity cardiovascular volumetric meshing

2025-06-09

AI Summary

Automated reconstruction of high-fidelity cardiovascular volumetric meshes from medical images remains a critical bottleneck for personalized physics-based simulation. Existing deep learning-based template deformation methods suffer from insufficient flexibility in high-curvature regions and geometric distortions in inter-component spacing. To address these limitations, we propose a “snap-and-tune” two-stage paradigm: first, coarse template registration via a U-Net architecture; second, unsupervised test-time optimization jointly minimizing mesh regularity and boundary alignment losses to refine local geometry. Our method requires no additional annotations, operates fully automatically, and produces meshes directly compatible with solid mechanics simulation. Experiments demonstrate significant improvements over baselines: a 37% reduction in Hausdorff distance, 99.2% inversion-free element rate, and successful deployment across multiple platforms for cardiac mechanical simulation incorporating realistic physiological boundary conditions.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Computer Vision: Medical and Biological Imaging
• Planning, Routing, and Scheduling: Replanning and Plan Repair

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments

Abstract

High-quality volumetric meshing from medical images is a key bottleneck for physics-based simulations in personalized medicine. For volumetric meshing of complex medical structures, recent studies have often utilized deep learning (DL)-based template deformation approaches to enable fast test-time generation with high spatial accuracy. However, these approaches still exhibit limitations, such as limited flexibility at high-curvature areas and unrealistic inter-part distances. In this study, we introduce a simple yet effective snap-and-tune strategy that sequentially applies DL and test-time optimization, which combines fast initial shape fitting with more detailed sample-specific mesh corrections. Our method provides significant improvements in both spatial accuracy and mesh quality, while being fully automated and requiring no additional training labels. Finally, we demonstrate the versatility and usefulness of our newly generated meshes via solid mechanics simulations in two different software platforms. Our code is available at https://github.com/danpak94/Deep-Cardiac-Volumetric-Mesh.

Problem

• High-quality volumetric meshing for medical images
• Limitations in flexibility at high-curvature areas
• Unrealistic inter-part distances in current methods

Innovation

• Combines deep learning with test-time optimization
• Automates high-fidelity volumetric meshing
• Improves spatial accuracy and mesh quality