Computed Tomography Image Reconstruction from Few-Views Data by Multi-Directional Total Variation

With the fact of harmfulness, X-ray radiation dose reduction received more and more attention in CT research field nowadays. As Compressed Sensing theory applied into Computed Tomography reconstruction algorithms make it possible to reconstruct high-quality CT images with the requirement of low radiation dose. One of research hot point in CS-based CT reconstruction algorithms is Total Variation, using the information of x-coordinate and y-coordinate gradients to reconstruct CT images, which implies that it will reconstruct better CT images by using gradients of more directions. In this paper, a CT reconstruction algorithm based on Multi-directional Total Variation (MdTV) was proposed, by using the information of x-coordinate, y-coordinate and diagonal gradient together to reconstruct CT images, which theoretically will reconstruct images with higher quality. Experimental results shows that, MdTV based reconstruction algorithm reconstruct better images than TV based algorithm, which is more helpful to reduce X-ray radiation dose as much as possible.