A Post-Processed Multi-Material Decomposition Method based on Dynamic Dual-energy Detectors

摘要

Due to the limitations of the data noise and spectrum distortion, accurate multi-material decomposition is a hard problem in spectral CT imaging research. Our group promoted a dynamic dual-energy photon counting detector. Its energy threshold can be dynamically changed at various energy bins during a spectral CT scan. It can provide more spectral information of the materials while keeping the data/image noise in a reasonable level. In this paper, we studied a post-processed multi-material-decomposition method based on this dynamic dual-energy detectors and tensor PRISM reconstruction method. We compared the reconstruction and decomposition results between the traditional static multi-energy CT and this dynamic dual-energy CT. The reconstruction results illustrated that the dynamic dual-energy CT can suppress the noise to a great extent compared to traditional static spectral CT, and can have similar results to reference true value. In the comparisons of decomposition results, some tissues are completely misclassified in the static spectral CT results while the dynamic dual-energy CT results have nearly the same results as the reference true value. Therefore, we conclude that the dynamic dual-energy CT can accurately decompose materials in similar components and have highly feasible in clinical practice.