GR-Gaussian: Graph-Based Radiative Gaussian Splatting for Sparse-View CT Reconstruction

2025-08-04

AI Summary

In sparse-view CT reconstruction, 3D Gaussian Splatting (3DGS) suffers from severe needle-like artifacts due to its reliance on intra-view average gradient magnitude. To address this, we propose a graph-structure-enhanced 3D Gaussian lattice rendering framework. Our method introduces: (1) a denoised point cloud initialization strategy to accelerate convergence; (2) a pixel-graph joint-aware gradient computation mechanism, leveraging graph neural networks to model voxel density correlations and enable density-difference-driven gradient optimization; and (3) synergistic optimization integrating pixel-level rendering with graph-structured priors. Evaluated on the X-3D and real-world sparse CT datasets, our approach achieves PSNR gains of 0.67–0.92 dB and SSIM improvements of 0.011–0.021 over baselines. It significantly suppresses artifacts while enhancing density representation accuracy and reconstruction fidelity.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Graph-based Machine Learning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

3D Gaussian Splatting (3DGS) has emerged as a promising approach for CT reconstruction. However, existing methods rely on the average gradient magnitude of points within the view, often leading to severe needle-like artifacts under sparse-view conditions. To address this challenge, we propose GR-Gaussian, a graph-based 3D Gaussian Splatting framework that suppresses needle-like artifacts and improves reconstruction accuracy under sparse-view conditions. Our framework introduces two key innovations: (1) a Denoised Point Cloud Initialization Strategy that reduces initialization errors and accelerates convergence; and (2) a Pixel-Graph-Aware Gradient Strategy that refines gradient computation using graph-based density differences, improving splitting accuracy and density representation. Experiments on X-3D and real-world datasets validate the effectiveness of GR-Gaussian, achieving PSNR improvements of 0.67 dB and 0.92 dB, and SSIM gains of 0.011 and 0.021. These results highlight the applicability of GR-Gaussian for accurate CT reconstruction under challenging sparse-view conditions.

Problem

• Reduces needle-like artifacts in sparse-view CT reconstruction
• Improves 3D Gaussian Splatting accuracy with graph-based methods
• Enhances density representation and convergence in CT imaging

Innovation

• Graph-based 3D Gaussian Splatting framework
• Denoised Point Cloud Initialization Strategy
• Pixel-Graph-Aware Gradient Strategy