Splat-Based Metal Artifact Reduction in Cone-Beam CT via Compact Attenuation Modeling

2026-08-05

AI Summary

This work addresses severe artifacts in cone-beam CT caused by high-attenuation metallic objects, which arise from the mismatch between the polychromatic nature of X-rays and the conventional monochromatic reconstruction assumption. The authors propose a splat-based metal artifact reduction framework that integrates a physics-driven polychromatic forward model into a continuous Gaussian scene representation. By employing a compact material parameterization, the method jointly optimizes geometric and material properties without requiring a metal mask, explicitly modeling energy-dependent attenuation. This approach effectively mitigates common limitations of existing techniques—such as smoothing bias, loss of fine details, and high computational cost—and demonstrates superior performance over both traditional reconstruction algorithms and neural radiance field methods on both simulated and real data, achieving faster convergence, enhanced artifact suppression, and improved preservation of high-frequency structural details.

Abstract

X-ray computed tomography (CT) suffers from severe metal artifacts when high-attenuation objects such as dental fillings or orthopedic implants are present. These artifacts originate from the polychromatic nature of X-rays, where attenuation varies strongly with photon energy and material composition, breaking the monochromatic assumption used by conventional reconstruction algorithms. Recent neural rendering approaches attempt to address this mismatch through differentiable polychromatic projection models, but they still struggle with smoothness bias, loss of fine structures, and prohibitive computation when extended to large-scale cone-beam CT. We introduce a splat-based metal artifact reduction framework that incorporates a physically grounded polychromatic forward model into a continuous Gaussian representation for cone-beam CT. Each Gaussian encodes the energy-dependent attenuation of the underlying material using a compact material parameterization, which enables efficient joint optimization of geometric and material properties without relying on a metal mask. This compact attenuation formulation captures the essential variation across biological tissues and metallic implants, allowing our model to explain metal-induced nonlinearity while preserving high-frequency structure. Experiments on simulated and real cone-beam CT scans show that our method converges significantly faster and suppresses metal artifacts more effectively than existing reconstruction and neural field-based approaches.

Problem

• metal artifact reduction
• cone-beam CT
• polychromatic X-ray
• attenuation modeling
• computed tomography

Innovation

• splat-based representation
• polychromatic forward model
• compact attenuation modeling
• metal artifact reduction
• cone-beam CT