Low dose CT image statistical reconstruction algorithms based on discrete shearlet

Reducing number of projection angles and lowering current intensity of X-ray tube are two common ways for reducing CT dose. Though reduced radiation dose of CT scan can lower damage to human bodies, Few number of projection angles will result in incomplete projection data while lowering tube current intensity a declined signal to noise ratio of projection data. In this paper, two statistical methods based on sparsity constraint in shearlet domain for low-dose CT image were proposed to solve the above problems. For the limited angle scanned reconstruction, sparse representation of intermediate images in shearlet domain is added into the objective function as a regularization item by means of Augmented Lagrangian method so as to narrow down solution space. For the low X-ray tube scanned reconstruction, a penalized weighted least-squares (PWLS) approach based on discrete shearlet was introduced to improve the performance of resisting noise in sinogram. And then reconstruct CT images by Filtered Back-Projection method. According to experimental data, both of the two approaches can get high-quality images when projection data is far from meeting conditions of completeness or the signal to noise ratio of projection data declines sharply. The proposed algorithms can be used for attaining reconstructed images that clearly keep structural details when the radiation dose is decreased to 10% or even lower degrees.