Multi-detector Computed Tomography offers the promise of a non-invasive alternative to invasive coronary angiography for the evaluation of coronary artery disease. An impediment preventing its widespread adoption is the presence of image "blooming" artifacts due to the presence vascular calcium. This blooming has been linked to cardiac motion, beam hardening, and resolution effects. In this paper we study the contribution of these elements to blooming in a controlled way and conclude that the strongest effect for current systems is due to resolution. We then present a multicomponent algebraic-type reconstruction approach to mitigate such blooming artifacts, motivated by recent results in image inpainting. The reconstruction approach decomposes the image into a collection of spatially localized components, each with a set of homogeneous properties. The local nature of the decomposition and constraints prevents artifacts from contaminating other image regions.
Zhuangli Liang, W. Clem Karl, Synho Do, Thomas Bra