A Model-Based Iterative Algorithm for Dual-Energy X-Ray CT Reconstruction

Recent developments in dual-energy X-ray CT have shown a number of benefits over standard CT for object separation, contrast enhancement, artifact reduction, and material composition assessment. As with traditional CT, model-based iterative approaches to reconstruction offer the opportunity to reduce noise and artifacts in dual energy reconstructions. However, previous approaches to model-based dual energy reconstruction have not fully modeled the statistical dependencies in the material-decomposed data. In this paper, we present a method for model-based iterative reconstruction which accounts for both the statistical dependency in the material decomposed sinogram components, and fast-switching approaches to dualenergy sampling. Our method also incorporates a positivity constraint in the space domain which accurately accounts for the true physical constraint of positive X-ray attenuation and is computationally simple to implement. Both phantom and clinical results show that the proposed model produces images which compare favorably to FBP in overall image quality.