Search papers, labs, and topics across Lattice.
This paper introduces a novel self-calibrating method for reducing metal artifacts in cone-beam computed tomography (CBCT) by integrating a polychromatic X-ray projection model with a Gaussian Splatting framework. The approach addresses the limitations of conventional monochromatic algorithms and existing neural representation methods, which are often computationally intensive and impractical for real-world applications. Experimental results show that this method significantly outperforms state-of-the-art techniques in both artifact suppression and reconstruction accuracy, while eliminating the need for manual metal masks or strong prior assumptions.
Efficiently reducing metal artifacts in CBCT without manual intervention, our method sets a new standard for reconstruction accuracy and practicality in medical imaging.
Cone鈥恇eam computed tomography (CBCT) enables volumetric reconstruction from X鈥恟ay projections, but suffers from severe artifacts鈥揺specially beam hardening鈥搘hen imaging materials with high attenuation such as metals. These artifacts arise from the polychromatic nature of X鈥恟ays and are not properly addressed by conventional monochromatic reconstruction algorithms. While recent neural representation鈥恇ased methods offer improved reconstruction quality, they are computationally expensive and often impractical for deployment. We propose a novel physics鈥恑nspired, self鈥恈alibrating metal artifact reduction method that efficiently reconstructs 3D CBCT volumes while correcting beam hardening artifacts. Our method integrates a polychromatic X鈥恟ay projection model, material鈥恉ependent attenuation profiles, and system response modeling into a Gaussian Splatting framework. Unlike prior work, we eliminate the need for manual metal masks or strong prior assumptions, and we optimize both reconstruction parameters and X鈥恟ay spectral characteristics jointly during training. We further introduce a high鈥恌idelity synthetic CBCT dataset generation pipeline validated on Monte鈥怌arlo x鈥恟ay simulation toolbox and release new datasets with severe metal鈥恑nduced artifacts to support the community. This is the first splat鈥恇ased method for reducing beam hardening in CBCT. Extensive experiments on both synthetic and real鈥恮orld datasets demonstrate that our method outperforms state鈥恛f鈥恡he鈥恆rt approaches in artifact suppression and reconstruction accuracy.
