Motion correction for CT angiography quality enhancement
Helen Hong, Kyoung Ho Lee, Yeong-Gil Shin · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2004
In this paper, we present a novel technique of improving vessel visualization quality by removing motion artifacts in digital subtraction brain CT angiography. The proposed methods based on the three key ideas as follows. First, the method involves the automatic selection of a set of feature points by using a 3D edge detection technique based on image gradient of mask and contrast volume. Second, locally weighted-3D distance map is generated to derive to robust convergence on the optimum value. Third, the similarity measure between extracted feature points is evaluated repeatedly by selective cross-correlation. The proposed method has been successfully applied to pre- and post-contrast CT angiography based on brain dataset for global and spatial motion correction. The feature point selection, introducing local processing on areas of interest consisting of voxels belonging to object boundary only, are very fast compared to all traditional algorithms where entire volume are searched. Since the registration estimates similarity measures between feature points and derive to robust convergence on the optimum value by the locally weighted-3D distance map, it offers an accelerated technique to accurately visualize vessels of the brain.