4DRGS: 4D Radiative Gaussian Splatting for Efficient 3D Vessel Reconstruction from Sparse-View Dynamic DSA Images

2024-12-17

arXiv.org

AI Summary

Reconstructing 3D vascular structures from sparse-view dynamic DSA images faces challenges of low accuracy, high computational cost, and difficulty balancing low-dose radiation requirements with clinical real-time constraints. Method: We propose a 4D radial Gaussian point-based representation: geometrically modeling the static vascular skeleton while dynamically encoding time-varying X-ray attenuation; incorporating cumulative attenuation pruning and bounded scaling activation to enhance reconstruction robustness; and integrating X-ray rasterization rendering, lightweight neural networks for temporal attenuation prediction, and voxel-based post-processing. Results: Evaluated on real patient data, our method trains in only five minutes and operates 32× faster than state-of-the-art approaches, while significantly improving reconstruction fidelity and clinical deployability under low-dose conditions.

Technology Category

• Computer Vision: 3D Computer Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Reconstructing 3D vessel structures from sparse-view dynamic digital subtraction angiography (DSA) images enables accurate medical assessment while reducing radiation exposure. Existing methods often produce suboptimal results or require excessive computation time. In this work, we propose 4D radiative Gaussian splatting (4DRGS) to achieve high-quality reconstruction efficiently. In detail, we represent the vessels with 4D radiative Gaussian kernels. Each kernel has time-invariant geometry parameters, including position, rotation, and scale, to model static vessel structures. The time-dependent central attenuation of each kernel is predicted from a compact neural network to capture the temporal varying response of contrast agent flow. We splat these Gaussian kernels to synthesize DSA images via X-ray rasterization and optimize the model with real captured ones. The final 3D vessel volume is voxelized from the well-trained kernels. Moreover, we introduce accumulated attenuation pruning and bounded scaling activation to improve reconstruction quality. Extensive experiments on real-world patient data demonstrate that 4DRGS achieves impressive results in 5 minutes training, which is 32x faster than the state-of-the-art method. This underscores the potential of 4DRGS for real-world clinics.

Problem

• Reconstruct 3D vessels from sparse-view dynamic DSA images
• Improve reconstruction quality and computation efficiency
• Model temporal contrast agent flow with 4D Gaussian kernels

Innovation

• 4D radiative Gaussian kernels for vessel modeling
• Neural network predicts time-dependent attenuation
• Accumulated attenuation pruning improves quality