Multispectral Limited-Angle Image Reconstruction

Limited-angle reconstruction algorithms utilize a priori information on the objects to compensate for the missing scannings in limited-angle imagings. In earlier papers we have shown qualitatively that decomposing a composite object through multiple energy scans enables better use of the a priori information and therefore is capable of bringing about improvement in limited-angle reconstructions. In this paper a more quantitative analysis is performed to study the conditions for such a improvement and its magnitude. In particular, it is shown that this technique favors reconstruction for the hot spot type of phantoms much more than the cold spot type.