Quantitative material decomposition with dynamic x-ray model in spectral CT

With the development of photon-counting detectors (PCD), spectral CT has gained greater flexibility in utilizing spectral information, making it a powerful tool for material decomposition. Traditional decomposition methods rely on predefined physical models to estimate effective atomic numbers and density. However, these models may fail to maintain accuracy under flexible scanning protocols or for a broader range of materials.

This study proposed an adaptive image-domain decomposition method (AIDM) to enhance the accuracy of material decomposition in spectral CT. AIDM dynamically adjusts the x-ray interaction model to compensate for modeling errors and improve quantitative material analysis.

To account for various physical effects, we introduce two correction terms: an exponential correction term, $C(E,Z_{\text{eff}})$ , that jointly models the photoelectric effect and coherent scattering, and a compensation factor, $D(E,Z_{\text{eff}})$ , for accurate Compton scattering correction. Both terms are formulated as polynomial expansions and fitted using NIST reference data. The effective energy for each bin is calibrated based on known reference materials, and decomposition is performed on reconstructed images using a nonlinear system derived from the improved attenuation model.

The proposed method was validated using standard materials, minerals, and biological samples. Compared to existing methods, AIDM demonstrated improved accuracy, robustness, and broader applicability in material quantification.

AIDM enables accurate spectral CT-based material estimation, providing significant improvements in various applications.