Improved C-arm cardiac cone beam CT based on alternate reconstruction and segmentation

摘要

Sparsity regularized iterative reconstruction is an important and promising method for ECG-gated tomo-graphic reconstruction of coronary artery during intervention treatment of cardiovascular diseases. As the reconstruction suffers from the problems of background overlay and data truncation, the background of angiogram should be well suppressed to obtain high reconstruction quality. Considering the deficiency of the commonly applied background suppression methods, this work proposes a strategy of alternate reconstruction and segmentation. During reconstruction, while the image intensity is iteratively updated, a contour is also evolved to segment the reconstructed vascular tree based on level set segmentation method. When the structure of the vascular tree is completely detected, the segmented vascular tree is re-projected to generate projection mask which is used to further reduce the projection background. Several experiments were performed to quantitatively evaluate the proposed method and the method is also compared with a state-of-the-art method. Experimental results show that the proposed strategy could effectively improve the reconstruction quality.