Generalized Gaussian model-based reconstruction method of computed tomography image from fewer projections

Abstract

In modern computed tomography (CT) imaging, a high-quality image is produced by passing the X-ray radiation dose through the human body. The strategy for CT dose reduction is to reduce the projection views or to limit the scan angle. However, the reduction of radiation doses generates insufficient projection data which often leads the reconstruction problem. Iterative techniques are widely used addressing this issue where a prior-based method is more beneficial to reconstruct CT images. In this paper, the soft-thresholding-based reconstruction method is presented, and the use of the generalized Gaussian distribution model is considered to find the optimal thresholding value. Experimental results show that the reconstructed CT image from fewer projection views using the proposed method has higher quality compared with the other reconstruction approaches.