Ordered-Subset Ridge Regression in Image Reconstruction

Weighted linear regression method has been widely applied for image reconstruction in computed tomography. In this paper we propose an approach based on the ridge regression method and derive two ordered-subset reconstruction algorithms from this framework. The convergence properties of the algorithms are validated by recent results on the ordered-subset Landweber scheme or numerical experiments. We also extend the strategy for formulating effective ordered-subset partitions based on the previous research studied by Jinhua Sheng. Numerical simulation and application to practical data demonstrate the performance of the proposed algorithms and ordered-subset strategy.