Affiliations: National Digital Switching System Engineering and
Technological R&D Centre, Zhengzhou, Henan, China
Note: [] Corresponding author: Bin Yan, National Digital Switching System
Engineering and Technological R&D Centre, Zhengzhou, Henan, China. E-mail:
[email protected]
Abstract: Linear scan computed tomography (CT) is a promising imaging
configuration with high scanning efficiency while the data set is under-sampled
and angularly limited for which high quality image reconstruction is
challenging. In this work, an edge guided total variation minimization
reconstruction (EGTVM) algorithm is developed in dealing with this problem. The
proposed method is modeled on the combination of total variation (TV)
regularization and iterative edge detection strategy. In the proposed method,
the edge weights of intermediate reconstructions are incorporated into the TV
objective function. The optimization is efficiently solved by applying
alternating direction method of multipliers. A prudential and conservative edge
detection strategy proposed in this paper can obtain the true edges while
restricting the errors within an acceptable degree. Based on the comparison on
both simulation studies and real CT data set reconstructions, EGTVM provides
comparable or even better quality compared to the non-edge guided
reconstruction and adaptive steepest descent-projection onto convex sets
method. With the utilization of weighted alternating direction TV minimization
and edge detection, EGTVM achieves fast and robust convergence and reconstructs
high quality image when applied in linear scan CT with under-sampled data
set.
Keywords: Linear scan computed tomography, limited angle problem, edge guided reconstruction, weighted total variation minimization, alternating direction method
