Algorithm-based low-dose computed tomography image reconstruction

Low-dose computed tomography (CT) reconstruction is a significant concern in CT imaging field. Currently, besides CT manufacturers adapted hardware techniques and optimized scan protocols to reduce the X-ray dose, algorithm-based low-dose CT reconstruction methods have been exploited extensively. However, for achieving high-quality algorithm-based low-dose CT reconstruction, there exist several challenges due to the excessive noise in low-dose projection data and the complex noise and artifacts characteristics in low-dose CT image. Statistical iterative reconstruction (SIR) methods have shown the potential to achieve a superior noise-resolution tradeoff as compared to analytical reconstruction techniques, however a main drawback of SIR is the computational burden associated with the multiple reprojection and back-projection operation cycles through the image domain. In this study, we propose an algorithm-based low-dose CT image reconstruction framework, which by making full use of the advantages of both the low-dose CT projection/sinogram data recovery and advanced edge-preserving CT image restoration. Simulated experimental results demonstrate that the present framework can yield image with better quality comparable to the obtained with the existing methods.