A Model of Regularization Parameter Determination in Low-Dose X-Ray CT Reconstruction Based on Dictionary Learning

Algorithm 1

Workflow of the developed algorithm.

Determination of

Initialize , and .

While the stopping criterion is not satisfied, do

 (1) Implement the OSC algorithm for acceleration;

 (2) Extract patches from the intermediate image ;

 (3) Update the dictionary by K-SVD algorithm;

 (4) Update the sparse coding by OMP algorithm;

 (5) Update the image by (25) with ;

Output the final image and relative error calculated by (26).

When the iteration is stopped, determine by (27).

Image Reconstruction

Initialize , and , is determined in former step.

While the stopping criterion is not satisfied, do

 (1) Implement the OSC algorithm for acceleration;

 (2) Extract patches from the intermediate image ;

 (3) Update the dictionary by K-SVD algorithm;

 (4) Update the sparse coding by OMP algorithm;

 (5) Update the image by (23), ;

Output the final reconstruction.