Application Research On Basis Material Decomposition Algorithm Of Dual-energy CT

In recent years, the clinical application of dual-energy CT (DECT) system is a significant technical progress in X-ray medical imaging. Compared with conventional CT, DECT is capable of differentiating tissue and eliminating beam hardening artifacts and plays a unique role in current diagnostic imaging. The research on DECT image reconstruction algorithm has become a hot spot in related fields.DECT can obtain the distribution of linear attenuation coefficients with two different X-ray spectra, reconstruct the monochromatic image, electron density image and effective atomic number image, and discriminate materials.But it is not easy for the researchers to acquire DECT raw projection data due to the limitation of intellectual property protection and insufficient technical support from manufacturers. And it has become a bottleneck for those theoretical explorations on DECT imaging. A method to realize the numerical simulation for the material decomposition and image reconstruction of DECT imaging is proposed in this thesis. The basic idea is to solve the linear equations and obtain the decomposition projections, which are used to reconstruct the monochromatic image, electron density image and effective atomic number image. The experimental results show the validity and the reliability of our method. We believe it is beneficial for those theoretical researchers of DECT imaging domain.Two material decomposition algorithm is mainly considered the basic materials pair, which consists of two kinds of materials with different linear attenuation properties. As we all know, it’s beneficial for the material discrimination if the linear attenuation property of the basis material is similar to the scanned subject. However, the complex composition of tissues and organs in the human body poses a challenge for many material decomposition methods, which assume the presence of only two, or at most three, materials in the mixture. So we implement an adaptive selection of optimal basis materials triplet method, which is based on the theory of three material decomposition model, by increasing the kinds of basis materials to obtain the better results of the discrimination. We use the geometric position relation of the decomposed pixels and basis materials triplet to select the optimal triplet, and then reconstruct the monochromatic image, electron density image and effective atomic number image. The experimental results of numerical simulation and those of clinical data show that the implemented method can effectively realize the image reconstruction of dual energy CT system, and reflect more detail scanning information. It will promote the development of dual energy CT imaging technology.