Nonlinear reconstruction of coded spectral X-ray CT based on material decomposition

Area of Science:

• Medical Imaging
• Computational Imaging
• Materials Science

Background:

• Coded spectral X-ray computed tomography (CT) offers cost-effective 3D structure and material composition analysis.
• Current methods face challenges in reconstructing spectral CT images from compressed, sparse measurements due to the nonlinear and ill-posed nature of the problem.

Purpose of the Study:

• To develop a novel material-decomposition-based approach for direct spectral CT image reconstruction.
• To overcome the ill-posedness and improve reconstruction quality using reduced dimensionality and advanced regularization techniques.

Main Methods:

• A material decomposition approach assuming linear attenuation coefficient maps are decomposable into basis materials.
• Reconstruction of mass density maps of basis materials, reducing optimization variable dimensionality.
• An alternating minimization scheme with weighted nuclear norm and total variation regularization.

Main Results:

• Significantly improved reconstruction quality compared to state-of-the-art methods for coded spectral CT.
• Successful reconstruction of spectral CT images at additional energy bins, enhancing spectral information retrieval.
• Effective reduction in inspection duration or radiation dose, crucial for biological imaging and medical diagnostics.

Conclusions:

• The proposed material decomposition method offers a robust solution for spectral CT image reconstruction from compressed measurements.
• This approach enhances diagnostic capabilities by providing richer spectral information and improving image fidelity.
• The method holds significant promise for advancing biological imaging and medical diagnostics through more efficient and accurate CT.