A multienergy computed tomography method without image segmentation or prior knowledge of X-ray spectra or materials

Area of Science:

• Medical Imaging
• Computed Tomography
• Image Reconstruction

Background:

• Traditional multienergy CT methods often require prior knowledge of X-ray spectra, materials, or image segmentation.
• These requirements can limit the applicability and efficiency of multienergy CT imaging.

Purpose of the Study:

• To propose a novel multienergy CT method that avoids the need for prior spectral or material information.
• To develop a blind decomposition model for accurate reconstruction of multienergy CT images.

Main Methods:

• A blind decomposition model is constructed to decompose multivoltage projections into energy-dependent coefficients and spatially dependent base images.
• The model utilizes variance minimization to address X-ray scattering, a source of low-frequency imaging errors.
• Karush-Kuhn-Tucker (KKT) conditions are employed to solve the decomposition model.

Main Results:

• The proposed method successfully synthesizes multienergy CT images from multivoltage projections acquired in a single scan.
• Reconstructed images demonstrate significantly reduced beam-hardening artifacts.
• The method achieves small relative errors in X-ray energy estimation for different materials, resulting in narrow energy intervals in the reconstructed images.

Conclusions:

• The developed blind decomposition method is effective for multienergy CT imaging.
• The technique offers a promising alternative for multienergy CT by eliminating the need for prior knowledge and reducing artifacts.