X-Field: A Physically Grounded Representation for 3D X-ray Reconstruction

2025-03-11

AI Summary

X-ray 3D reconstruction faces dual challenges: low-dose constraints and physical model mismatch. Existing methods adapt visible-light neural radiance fields (NeRFs), neglecting the fundamental X-ray attenuation physics—particularly energy-dependent material absorption—leading to inaccurate volumetric representation. To address this, we propose X-Field, the first X-ray–aware 3D volume representation grounded in first-principles X-ray physics. X-Field employs a 3D ellipsoidal parameterization for multi-material attenuation modeling, integrates physics-driven ray segmentation and hybrid progressive initialization, and introduces material-aware gradient optimization alongside a forward-projection–backprojection joint training framework. Evaluated on both real organ datasets and synthetic benchmarks, X-Field achieves state-of-the-art performance in novel-view synthesis and CT reconstruction, with significant improvements in visual fidelity and quantitative metrics (PSNR/SSIM). This demonstrates the critical importance of physically consistent modeling for low-dose X-ray imaging.

Technology Category

• Computer Vision: Low Level & Physics-based Vision
• Humans and AI: Explainable AI (XAI) for Human Understanding
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs

Abstract

X-ray imaging is indispensable in medical diagnostics, yet its use is tightly regulated due to potential health risks. To mitigate radiation exposure, recent research focuses on generating novel views from sparse inputs and reconstructing Computed Tomography (CT) volumes, borrowing representations from the 3D reconstruction area. However, these representations originally target visible light imaging that emphasizes reflection and scattering effects, while neglecting penetration and attenuation properties of X-ray imaging. In this paper, we introduce X-Field, the first 3D representation specifically designed for X-ray imaging, rooted in the energy absorption rates across different materials. To accurately model diverse materials within internal structures, we employ 3D ellipsoids with distinct attenuation coefficients. To estimate each material's energy absorption of X-rays, we devise an efficient path partitioning algorithm accounting for complex ellipsoid intersections. We further propose hybrid progressive initialization to refine the geometric accuracy of X-Filed and incorporate material-based optimization to enhance model fitting along material boundaries. Experiments show that X-Field achieves superior visual fidelity on both real-world human organ and synthetic object datasets, outperforming state-of-the-art methods in X-ray Novel View Synthesis and CT Reconstruction.

Problem

• Develops X-Field for 3D X-ray imaging representation.
• Models material energy absorption using 3D ellipsoids.
• Enhances X-ray view synthesis and CT reconstruction accuracy.

Innovation

• X-Field: 3D representation for X-ray imaging
• 3D ellipsoids model material attenuation coefficients
• Path partitioning algorithm for energy absorption estimation