Efficient TpV minimization for circular, cone-beam computed tomography reconstruction via non-convex optimization

Cone-beam computed tomography (CT) configured with circular scanning is widely used in many imaging applications [1]. Accurate image reconstruction with sparse or few views is greatly important in areas needing fast data acquisition [2] or low radiation levels [3]. A lot of work [2–13] regarding iterative image reconstruction (IIR) algorithms has focused on total variation (TV) minimization, which exploits the sparsity of the gradient magnitude image (GMI). TV is the L1 norm of the GMI and reconstruction done by minimizing TV can reduce data acquisition when the GMI is sparse in non-zero entities.
Source: Computerized Medical Imaging and Graphics - Category: Radiology Authors: Source Type: research